CN114937055A - Image self-adaptive segmentation method and system based on artificial intelligence - Google Patents
Image self-adaptive segmentation method and system based on artificial intelligence Download PDFInfo
- Publication number
- CN114937055A CN114937055A CN202210356859.0A CN202210356859A CN114937055A CN 114937055 A CN114937055 A CN 114937055A CN 202210356859 A CN202210356859 A CN 202210356859A CN 114937055 A CN114937055 A CN 114937055A
- Authority
- CN
- China
- Prior art keywords
- gray
- image
- pixel point
- threshold
- gray level
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 230000011218 segmentation Effects 0.000 title claims abstract description 92
- 238000000034 method Methods 0.000 title claims abstract description 50
- 238000013473 artificial intelligence Methods 0.000 title claims abstract description 22
- 238000012545 processing Methods 0.000 claims abstract description 16
- 230000007547 defect Effects 0.000 claims description 130
- 230000003044 adaptive effect Effects 0.000 claims description 16
- 230000002950 deficient Effects 0.000 claims description 10
- 238000004458 analytical method Methods 0.000 claims description 7
- 238000007405 data analysis Methods 0.000 claims description 6
- 238000010276 construction Methods 0.000 claims 1
- 230000000694 effects Effects 0.000 abstract description 9
- 230000008569 process Effects 0.000 description 9
- 238000001514 detection method Methods 0.000 description 8
- 238000003709 image segmentation Methods 0.000 description 7
- 230000002902 bimodal effect Effects 0.000 description 5
- 238000012937 correction Methods 0.000 description 5
- 230000008859 change Effects 0.000 description 4
- 238000004364 calculation method Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 230000007613 environmental effect Effects 0.000 description 2
- 241000221535 Pucciniales Species 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000007797 corrosion Effects 0.000 description 1
- 238000005260 corrosion Methods 0.000 description 1
- 230000009977 dual effect Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 238000010191 image analysis Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000009659 non-destructive testing Methods 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000000926 separation method Methods 0.000 description 1
- 238000012549 training Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/136—Segmentation; Edge detection involving thresholding
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/40—Image enhancement or restoration using histogram techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/60—Analysis of geometric attributes
- G06T7/62—Analysis of geometric attributes of area, perimeter, diameter or volume
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/90—Determination of colour characteristics
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20004—Adaptive image processing
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Geometry (AREA)
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The invention relates to the technical field of artificial intelligence, in particular to an image self-adaptive segmentation method and system based on artificial intelligence, wherein the method comprises the following steps: acquiring a target gray image, counting the gray value of each pixel point on the image, constructing a gray histogram, and calculating the probability of each gray level based on the gray histogram; performing sliding window processing on the target gray level image, and calculating the chaos degree of a central pixel point in the sliding window according to the frequency of the pixel point in the sliding window corresponding to the gray level and the variance of the gray level; correcting the gray value of the corresponding pixel point by utilizing the chaos degree to obtain the gray value of the corrected pixel point, and establishing a corrected gray histogram, wherein the corrected histogram has obvious double-peak characteristics; and determining a segmentation threshold value based on the double-peak characteristics of the corrected gray level histogram, and performing threshold segmentation on the target gray level image. The method has better segmentation effect when the threshold segmentation is carried out on the image without obvious double peak characteristics.
Description
Technical Field
The invention relates to the technical field of artificial intelligence, in particular to an image self-adaptive segmentation method and system based on artificial intelligence.
Background
Image segmentation is an important step in the analysis of images, and among all image segmentation algorithms, threshold segmentation has been widely used for a long time due to its extremely compact and highly practical characteristics. The image segmentation links the bottom layer processing and the high layer processing together, and aims to divide the image into a set of non-overlapping regions corresponding to the real world, so that the separation of an interested target and other regions in the image is realized, and the interested target can be further processed by adopting high-level visual technologies such as tracking, detection, identification and the like.
The threshold segmentation method is widely applied to image analysis application occasions where the gray levels of pixels of a target and a background have obvious differences, and typical applications include: extracting characters and marks in the text image; extracting the identifier in the map image; extracting a target in a scene image; extracting a defect part mark in the material quality detection process; cell image processing, thermal image processing, non-destructive testing, video spatiotemporal image segmentation, and the like. The basic idea of thresholding is to determine a threshold between the minimum and maximum grey levels of an image and then to classify all pixels in the image into two classes with their grey levels bounded by the threshold.
The traditional threshold segmentation algorithm has a good segmentation effect on images with obvious double-peak characteristics of histograms and small difference of target and background variances. However, in the image acquisition process, the images are often influenced by various environmental interference factors, so that the acquired images do not have obvious bimodal features, and meanwhile, partial images do not have bimodal features, and a traditional threshold segmentation algorithm is difficult to use in the situation. If the segmentation threshold is determined by manual marking, the workload is large, and the generalization capability is weak.
Disclosure of Invention
In order to solve the above technical problems, an object of the present invention is to provide an image adaptive segmentation method based on artificial intelligence, which adopts the following technical solutions:
step 1, acquiring a target gray image, counting gray values of all pixel points on the image, constructing a gray histogram, and calculating the probability of each gray level based on the gray histogram;
step 2, performing sliding window processing on the target gray level image, and calculating the chaos degree of a central pixel point in the sliding window according to the frequency of the gray level corresponding to the pixel point in the sliding window and the variance of the gray level, so as to obtain the chaos degree of all the pixel points on the target gray level image;
step 3, correcting the gray value of the corresponding pixel point by utilizing the chaos degree to obtain the gray value of the corrected pixel point, and establishing a corrected gray histogram, wherein the corrected histogram has obvious double-peak characteristics;
and 4, determining a segmentation threshold value based on the double-peak characteristics of the modified gray level histogram, and performing threshold segmentation on the target gray level image.
Preferably, before constructing the gray level histogram, the method further includes a step of extracting a defect region from the target gray level image, specifically: setting a prior threshold and carrying out threshold segmentation on the target gray level image to obtain a defective area and a non-defective area.
Preferably, judging the size of the area of the defect region and a set defect area threshold, and if the area of the defect region is smaller than the set defect area threshold, performing threshold segmentation of the step 1 to the step 4 on the target gray level image; if the area of the defect region is larger than the set defect area threshold, the threshold segmentation is not performed.
Preferably, the method for obtaining the chaos of the pixel point specifically comprises the following steps:
wherein Q a Expressing the chaos of the pixel point a, the size of a sliding window with the pixel point a as the center is n multiplied by n, p (X) expresses the probability when the gray level of the pixel point in the sliding window is X, X expresses the set of gray level series of each pixel point in the sliding window, n 2 -1 represents the number of pixels within the sliding window except the center pixel, S i Representing the gray level of the ith pixel point in the sliding window,
representing images in sliding windowsMean of the grey levels of the pixels.
Preferably, the step of extracting the defective area of the target grayscale image specifically includes: the part of the gray value of the pixel point on the target gray image, which is smaller than the prior threshold value, is a defect area, and the part of the gray value of the pixel point on the target gray image, which is larger than the prior threshold value, is a non-defect area.
The invention also provides an image self-adaptive segmentation system based on artificial intelligence, which comprises:
the data acquisition module is used for acquiring a target gray image, counting the gray value of each pixel point on the image, constructing a gray histogram and calculating the probability of each gray level based on the gray histogram;
the data analysis module is used for performing sliding window processing on the target gray level image, calculating the chaos degree of a central pixel point in the sliding window according to the frequency of the occurrence of the gray level corresponding to the pixel point in the sliding window and the variance of the gray level, and further acquiring the chaos degree of all the pixel points on the target gray level image;
the threshold segmentation module is used for correcting the gray value of the corresponding pixel point by utilizing the chaos degree to obtain the gray value of the corrected pixel point and establishing a corrected gray histogram, wherein the corrected histogram has obvious double-peak characteristics; and determining a segmentation threshold value based on the double-peak characteristics of the modified gray level histogram, and performing threshold segmentation on the target gray level image.
Preferably, the data acquisition module further includes a defect region analysis unit, configured to set a prior threshold and perform threshold segmentation on the target grayscale image to obtain a defect region and a non-defect region.
Preferably, the defect area analysis unit is further configured to determine the size of the defect area and a set defect area threshold, and if the defect area is smaller than the set defect area threshold, input the target grayscale image into the data analysis module for subsequent operations; and if the area of the defect area is larger than the set defect area threshold value, outputting a target gray level image.
Preferably, the method for obtaining the chaos of the pixel point specifically comprises the following steps:
wherein Q is a Expressing the chaos of the pixel point a, the size of a sliding window with the pixel point a as the center is n multiplied by n, p (k) expresses the probability when the gray level of the pixel point in the sliding window is X, X expresses the set of gray level series of each pixel point in the sliding window, n 2 -1 represents the number of pixels in the sliding window except the center pixel, S i Representing the gray level of the ith pixel point in the sliding window,
and expressing the average value of the gray levels of all the pixel points in the sliding window.
Preferably, the step of extracting the defective area of the target grayscale image specifically includes: the part of the gray value of the pixel point on the target gray image, which is smaller than the prior threshold value, is a defect area, and the part of the gray value of the pixel point on the target gray image, which is larger than the prior threshold value, is a non-defect area.
The embodiment of the invention at least has the following beneficial effects:
the method calculates the chaos degree of the pixel point according to the probability of the occurrence of the gray level of the pixel in the neighborhood of the pixel point and the gray level variance, corrects the gray level of the pixel point according to the chaos degree of the pixel point, further obtains a gray level histogram with obvious double-peak characteristics after correction, and obtains an optimal segmentation threshold value for threshold segmentation. The gray level histogram correction method not only considers gray level characteristics, but also weights the gray level values by combining the spatial information of the image to obtain the corrected gray level histogram, has obvious double-peak characteristics, and has better segmentation effect when the threshold segmentation is carried out on the image without the obvious double-peak characteristics.
Meanwhile, the method also comprises the step of extracting the defect region of the target gray image before constructing the gray histogram, and whether the threshold segmentation of the gray image is carried out is determined by judging the area size of the defect region. The method comprises the steps of carrying out pre-judgment before self-adaptive threshold segmentation to determine the defect size of a target gray image, wherein when a small-area defect occurs, the change of a pixel value of a defect area is not obvious, the contained characteristic information is relatively less, and the defect information is easy to ignore during detection, so that the conventional threshold segmentation cannot be accurately segmented.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions and advantages of the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and other drawings can be obtained by those skilled in the art without creative efforts.
Fig. 1 is a flowchart of a method for image adaptive segmentation based on artificial intelligence according to embodiment 1 of the present invention;
fig. 2 is a system block diagram of an artificial intelligence based image adaptive segmentation system according to embodiment 2 of the present invention.
Detailed Description
To further illustrate the technical means and effects of the present invention adopted to achieve the predetermined objects, the following detailed description will be given to an image adaptive segmentation method and system based on artificial intelligence according to the present invention, with reference to the accompanying drawings and preferred embodiments, and the detailed implementation, structure, features and effects thereof. In the following description, different "one embodiment" or "another embodiment" refers to not necessarily the same embodiment. Furthermore, the particular features, structures, or characteristics may be combined in any suitable manner in one or more embodiments.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.
The following describes a specific scheme of an image adaptive segmentation method and system based on artificial intelligence in detail with reference to the accompanying drawings.
Example 1:
referring to fig. 1, a flowchart illustrating steps of an artificial intelligence-based image adaptive segmentation method according to an embodiment of the present invention is shown, where the method includes the following steps:
step 1, obtaining a target gray image, counting gray values of all pixel points on the image, constructing a gray histogram, and calculating the probability of each gray level based on the gray histogram.
Specifically, a product surface image is acquired, the acquired image is preprocessed, and the processed image is converted into a gray image serving as a target gray image. The method comprises the steps of collecting a product surface image by a camera, preprocessing the obtained image, removing the influence of other environmental interference and noise, and converting the processed image into a gray image serving as a target gray image. In the embodiment, the mean filtering denoising algorithm is adopted to preprocess the acquired image, and an implementer can select other denoising algorithms to preprocess the acquired image according to actual conditions, so as to remove noise influence in the image acquisition process.
It should be noted that, for an image with obvious double-peak features in the gray histogram, the difference between the target and the background is obvious, and the result is ideal when the gray histogram is used to determine the segmentation threshold and further perform image segmentation. However, for an image with low quality or an image with an insignificant difference between an object and a background, the threshold segmentation method using the histogram dual peak method has a poor effect. The traditional histogram bimodal method only considers the gray characteristic of an image in threshold segmentation, the invention combines the spatial domain information of the image for weighting, and considers the chaos degree around each pixel point on the image, thereby increasing the difference between a target and a background in the image, enabling the target and the background to present the bimodal characteristic of the histogram, and selecting the image segmentation threshold so as to achieve the purpose of self-adaptive threshold segmentation.
Specifically, the gray value of each pixel point on the target gray image is counted, a gray histogram is established, the frequency of each gray level is calculated, and the formula is expressed as follows:
wherein, P k Representing the probability of a gray level of k, A k And B represents the total number of pixel points on the target gray level image.
In the target gray image, the gray level difference between the target area and the background area is not large, so that the gray level histogram of the image may be in a single peak state or an unobvious double peak state, and the background area corresponds to the first peak point. The gray-scale histogram may also have a plurality of peak values due to interference of various factors, and there is a large interference on the selection of the threshold, so that the gray-scale value needs to be weighted by combining spatial information of the image, so as to modify the gray-scale histogram and amplify the difference between the second peak value and other pseudo peak values.
And 2, performing sliding window processing on the target gray level image, and calculating the chaos degree of a central pixel point in the sliding window according to the frequency of the gray level corresponding to the pixel point in the sliding window and the variance of the gray level, so as to obtain the chaos degree of all the pixel points on the target gray level image.
Specifically, the chaos of the pixel points is expressed by a formula as follows:
wherein Q a Expressing the chaos of the pixel point a, the size of a sliding window with the pixel point a as the center is n multiplied by n, p (X) expresses the probability when the gray level of the pixel point in the sliding window is X, X expresses the set of gray level series of each pixel point in the sliding window, n 2 -1 represents the number of pixels within the sliding window except the center pixel, S i Representing the gray level of the ith pixel point in the sliding window,
and expressing the average value of the gray levels of all the pixel points in the sliding window.
For the background area, the gray values of the partial pixel points and the neighborhood pixel points are similar, so that the gray values of the pixel points are not different under the general condition, and even if the difference exists, the difference of the gray values is smaller, so that the information entropy of the pixel points in the background area is small, the contained information quantity is small, the gray variance is small, the gray fluctuation is small, and the disorder degree is small.
For the target area, because the gray values of the pixel points included in the target area are different, and the gray values of the pixel points are different due to various influence factors, the gray values of the pixel points in the target area and the pixel points in the neighborhood are different, the information entropy of the pixel points is large, the contained information amount is large, the gray variance is large, the gray fluctuation is large, and the degree of confusion is large.
For the edge pixel points of the target area and the background area, the gray values of the pixel points in the neighborhood are often greatly different. Because the pixel points are edge pixel points, and the neighborhood pixel points comprise the pixel points in the target area and the background area, the gray values of the edge pixel points and the neighborhood pixel points are often in a sudden change rule, so that the information entropy of the edge pixel points is large, the contained information is large, the gray variance is large, the gray fluctuation is large, and the chaos is large.
Step 3, correcting the gray value of the corresponding pixel point by utilizing the chaos degree to obtain the gray value of the corrected pixel point, and establishing a corrected gray histogram; the modified histogram has a pronounced bimodal character.
Specifically, the gray value of the corresponding pixel point is corrected by using the chaos degree of the pixel point, and the gray value is expressed by a formula:
I a ＝(1-Q a )*Z a
wherein, I a Representing the corrected gray value, Q, of the pixel point a a Indicating the degree of disorder, Z, of pixel a a And representing the gray value of the pixel point a before correction.
Because the chaos of the pixel points in the background area is small, the influence of the chaos on the gray value of the pixel points after being corrected is small, the gray value of the pixel points after being corrected is similar to the gray value before being corrected, correspondingly, the gray level of the pixel points after being corrected is similar to the gray level before being corrected, the number of the pixel points corresponding to the gray levels before and after being corrected is also similar, and the peak value of the original gray level histogram related to the background area can not be influenced after being corrected.
For the pixel points in the target area, the gray values of the pixel points are corrected by utilizing the chaos degree, and then the difference between the pixel points in the target area and the pixel points in the background area can be increased. The gray value of the pixel point similar to that in the background area may exist in the target area, and after the chaos degree is corrected, the difference between the gray value of the pixel point and the gray value of the pixel point in the background area is increased, so that a plurality of local small peak points are fused and recombined into one peak point, and the corresponding gray histogram has obvious double-peak characteristics.
And 4, determining a segmentation threshold value based on the double-peak characteristics of the modified gray level histogram, and performing threshold segmentation on the target gray level image.
Specifically, an optimal segmentation threshold is calculated between two peaks on the corrected gray level histogram, and segmentation processing is performed according to the segmentation threshold to obtain a target region and a background region. The optimal segmentation threshold is as follows: the optimal segmentation threshold is utilized to carry out threshold segmentation, so that the optimal segmentation effect can be achieved. For the segmentation effect, the optimal segmentation effect is that the intra-class variance is minimum, so a value is selected between double peaks on the corrected gray level histogram for segmentation, and the intra-class variance is calculated. The calculation of the intra-class variance is performed in the prior art, and is not repeated here, different values between two peaks are selected for repeated calculation, and the gray value corresponding to the value with the minimum intra-class variance is selected, and the gray value is the optimal segmentation threshold.
Example 2:
the image adaptive segmentation method based on artificial intelligence provided by the embodiment is different from the method provided by the embodiment 1 only in that: before constructing the gray level histogram, the method also comprises the steps of extracting the defect area of the target gray level image, and judging the area of the defect area and setting the size of the defect area threshold.
In the embodiment, a target gray image containing a defect is taken as an example, a roller large end surface gray image of a roller bearing containing the defect is obtained, the obtained image is processed, the influence of noise in the image acquisition process is removed, the processed roller large end surface gray image is taken as the target gray image, the target gray image is pre-segmented by adopting a traditional threshold segmentation algorithm, the area of a defect region and the size of a set defect area threshold are judged, the image with large-area defects is eliminated, and if the gray feature of the image with small area of the screened defect region is not obvious, the target gray image is segmented by adopting an adaptive threshold segmentation method to obtain the defect region.
Specifically, the method provided by this embodiment includes the following steps:
firstly, acquiring a roller large end surface image of a roller bearing, processing the acquired image, and graying to obtain a target grayscale image.
Specifically, a camera light source and a background plate are arranged, the camera is located right above the large end face of the roller bearing, the large end face surface image of the roller bearing is collected in a overlooking mode, the large end face of the roller bearing is located on the white background plate and is used for preventing the interference of a complex environment on subsequent image segmentation, the light source is an LED white light bar-shaped light source, and the light source is located obliquely above a product.
And processing the collected roller large end surface image of the roller bearing, removing the influence of background environment interference and noise, and graying the processed image to obtain a target grayscale image. In this embodiment, a DNN network is used to process an image of a large end surface of a roller bearing, and the specific method is as follows:
the DNN network data set is a roller large end surface image data set of the roller bearing acquired in the acquisition process; the roller bearing has various roller end surface patterns. The pixel points to be divided are of two types, namely the label marking process of the reading and research of the training set is as follows: the semantic label of the single channel is that the pixel point at the corresponding position belongs to the background area and is marked as 0, and the pixel point at the corresponding position belongs to the large end face of the roller bearing and is marked as 1; the task of the network is classification, and all used loss functions are cross entropy loss functions.
It should be noted that the implementer may select another suitable method such as semantic segmentation to remove the interference of the background area on the image according to the actual situation.
Then setting a prior threshold and carrying out threshold segmentation on the target gray level image to obtain a defective area and a non-defective area; the part of the gray value of the pixel point on the target gray image, which is smaller than the prior threshold value, is a defect area, and the part of the gray value of the pixel point on the target gray image, which is larger than the prior threshold value, is a non-defect area.
Specifically, the large-area defects on the surface of the product are mainly caused by factors such as collision between workpieces or insufficient generation processes, and the defects present more obvious gray-scale features in the image of the surface of the product. When the product quality with obvious defect information is detected, the image is segmented according to the prior set prior threshold value to obtain a defect region and a non-defect region. According to the fact that the gray value of the defect area with a large prior area is lower than that of the non-defect area in a normal condition, the part of the target gray image with the gray value of the pixel point smaller than the set threshold is the defect area, and the part of the target gray image with the gray value of the pixel point larger than the set threshold is the non-defect area, and the defect area can be expressed as follows:
wherein g (x, y) represents the gray scale value of the pixel point at the (x, y) position after the threshold segmentation, f (x, y) represents the gray scale value of the pixel point at the (x, y) position before the threshold segmentation, (x, y) represents the coordinates of the pixel point on the target gray image, T represents the prior threshold set according to the prior, and the value of the prior threshold is T ═ 100 in this embodiment.
Selecting a set defect area threshold according to prior, judging the size of the area of the defect area and the set defect area threshold, and if the area of the defect area is smaller than the set defect area threshold, performing threshold segmentation operation of steps 1-4 in the method provided by embodiment 1 on the target gray level image; if the area of the defect area is larger than the set defect area threshold, the threshold segmentation is not carried out. The area threshold is set according to different threshold dividing objects and actual conditions, and in the embodiment, the area threshold is 20% of the area of the large end face of the roller bearing, and the operator can set the area threshold according to the actual conditions
Specifically, if the area of the defect area is larger than the set defect area threshold, the large-area defect is contained in the large roller end face of the roller bearing corresponding to the target gray image, which indicates that the roller bearing is unqualified in quality, and the threshold segmentation operation is not needed to be performed on the target gray image. If the area of the defect region is smaller than the area threshold, the large-area defect does not exist in the large end face of the roller bearing corresponding to the target gray-scale image, and the defect may be small or free, so that the target gray-scale image needs to be subjected to adaptive segmentation to obtain the defect region.
In the product surface defect detection, since the characteristic information contained in the product surface image with a large area of defect is obvious, the detection is relatively high in accuracy and easy to detect, while the characteristic information contained in the image with a small defect is relatively less and easy to ignore in detection, in order to increase the accuracy of the defect detection, the product surface image with a large area of defect needs to be filtered out first.
Further, if the defect area is smaller than the set defect area threshold, the threshold segmentation operation from step 1 to step 4 in the method provided in embodiment 1 is performed on the target gray-scale image. This step is different from embodiment 1 in that the target is a defective portion and the background is a normal portion in this embodiment.
The bearing roller has many surface defects, which mainly include damage, scratches, cracks caused by mechanical impact, and rusty spots, missing materials, stains, etc. caused by the roller during the operation of other devices. Since these defects are reflected in the image in a state of excessively low luminance or excessively high luminance, it is presumed that the probability of the defects appearing on the left and right sides of the gradation histogram is high and the number of the defects appearing in the center area of the gradation histogram is almost 0. Because the proportion of the defects is far smaller than that of the normal part, the gray level with the highest occurrence probability is the gray level of the normal part, and the gray level with the highest occurrence probability is selected as the gray level of the normal part on the surface of the large end face of the roller bearing.
Because the gray scale difference between the defect part and the normal part in the target gray scale image is not large, the gray scale histogram of the image may be in a single peak state or an unobvious double peak state, and the normal part corresponds to the first peak point. The gray level histogram may have a plurality of peak values due to interference of various factors, and there is a large interference on the selection of the threshold, so that the chaos degree of the pixel points needs to be calculated by combining the spatial information of the image, and the gray level of the pixel points is weighted by using the chaos degree of the pixel points, so that the gray level histogram is corrected, and the difference between the second peak value and other pseudo peak value points is amplified.
For a normal part, the gray values of the part of pixel points are similar to the gray values of the neighborhood pixel points, so that the gray values of the pixel points are not different under the general condition, and even if the difference exists, the difference of the gray values is smaller, so that the information entropy of the normal part of the pixel points is small, the contained information quantity is small, the gray variance is small, the gray fluctuation is small, and the disorder degree is small.
The defective portion is a defect caused by various factors, such as damage, scratches, cracks caused by mechanical impact, and rusts, chips, stains, and the like caused by the roller during the operation of other equipment. Compared with the adjacent pixels, the pixels of the defect part have different reasons (such as stress degree, corrosion degree and the like) which cause the defect, so that the gray difference of the defect part often exists, the gray values of the pixels of the defect part and the adjacent pixels often show a gradual change rule, the information entropy of the pixels of the defect part is large, the contained information is large, the gray variance is large, the gray fluctuation is large, and the chaos is large.
For edge pixel points for distinguishing defects from non-defects, the gray values of the edge pixel points and the neighboring pixel points are often greatly different. Because the pixel points are edge pixel points, and the neighborhood pixel points comprise defect part pixel points and normal part pixel points, the gray values of the edge pixel points and the neighborhood pixel points are often in a sudden change rule, so that the information entropy of the edge pixel points is large, the contained information is large, the gray variance is large, the gray fluctuation is large, and the chaos degree is large.
Because only the gray level histogram considering the gray level information presents a single peak state, the gray level of the corresponding pixel point is corrected by utilizing the chaos degree to obtain the gray level of the corrected pixel point, and the corrected gray level histogram is established; and determining a segmentation threshold value by using the corrected gray level histogram, and performing threshold segmentation on the target gray level image.
The confusion degree of the pixel points of the normal part is small, the influence of the confusion degree on the gray value of the pixel points is small, the gray value of the pixel points after being corrected is similar to the gray value before being corrected, correspondingly, the gray level of the pixel points after being corrected is similar to the gray level before being corrected, the number of the pixel points corresponding to the gray levels before and after being corrected is also similar, and the peak value of the normal part in the original gray histogram can not be influenced after being corrected.
For the pixel points of the defect part, the gray values of the pixel points are corrected by utilizing the chaos degree, and then the difference between the pixel points of the defect part and the pixel points of the normal part can be increased. The gray value of the pixel point similar to that of the normal part may exist in the defect part, after the chaos degree correction is performed, the difference between the gray value of the pixel point and the gray value of the pixel point of the normal part is increased, and the number of the pixel points of the defect part is only inferior to that of the normal part (in this embodiment, the pixel points are roughly classified into three types, namely, defect, non-defect, and edge), so that a plurality of small local peak points are fused and recombined into one peak point after the correction, and further, the corresponding gray histogram has a relatively obvious double peak characteristic.
And establishing a modified gray histogram according to the modified gray value of the pixel point, calculating an optimal segmentation threshold, and performing segmentation processing according to the segmentation threshold to obtain a defect part and a normal part so as to obtain a detection result of the surface defect of the large end face of the roller bearing.
Example 3:
referring to fig. 2, a block diagram of a system for an artificial intelligence based image adaptive segmentation system according to an embodiment of the present invention is shown, where the system includes:
the data acquisition module is used for acquiring a target gray image, counting the gray value of each pixel point on the image, constructing a gray histogram and calculating the probability of each gray level based on the gray histogram;
the data analysis module is used for performing sliding window processing on the target gray level image, calculating the chaos degree of a central pixel point in the sliding window according to the frequency of the occurrence of the gray level corresponding to the pixel point in the sliding window and the variance of the gray level, and further acquiring the chaos degree of all the pixel points on the target gray level image;
the threshold segmentation module is used for correcting the gray value of the corresponding pixel point by utilizing the chaos degree to obtain the gray value of the corrected pixel point and establishing a corrected gray histogram, wherein the corrected histogram has obvious double-peak characteristics; and determining a segmentation threshold value based on the double-peak characteristics of the corrected gray level histogram, and performing threshold segmentation on the target gray level image.
Preferably, the data acquisition module further includes a defect region analysis unit, configured to set a prior threshold and perform threshold segmentation on the target grayscale image to obtain a defect region and a non-defect region.
Preferably, the defect area analysis unit is further configured to determine the size of the defect area and a set defect area threshold, and if the defect area is smaller than the set defect area threshold, input the target grayscale image into the data analysis module for subsequent operations; and if the area of the defect area is larger than the set defect area threshold value, outputting a target gray level image.
Preferably, the method for obtaining the chaos of the pixel points specifically comprises the following steps:
wherein Q is a Expressing the chaos of the pixel point a, the size of a sliding window with the pixel point a as the center is n multiplied by n, p (X) expresses the probability when the gray level of the pixel point in the sliding window is X, and X expresses the gray level number of each pixel point in the sliding windowSet of (2), n 2 -1 represents the number of pixels in the sliding window except the center pixel, S i Representing the gray level of the ith pixel point in the sliding window,
and representing the mean value of the gray levels of all pixel points in the sliding window.
Preferably, the step of extracting the defect area of the target grayscale image specifically includes: the part of the gray value of the pixel point on the target gray image, which is smaller than the prior threshold value, is a defect area, and the part of the gray value of the pixel point on the target gray image, which is larger than the prior threshold value, is a non-defect area.
It should be noted that: the precedence order of the above embodiments of the present invention is only for description, and does not represent the merits of the embodiments. And that specific embodiments have been described above. Other embodiments are within the scope of the following claims. In some cases, the actions or steps recited in the claims may be performed in a different order than in the embodiments and still achieve desirable results. In addition, the processes depicted in the accompanying figures do not necessarily require the particular order shown, or sequential order, to achieve desirable results. In some embodiments, multitasking and parallel processing may also be possible or may be advantageous.
All the embodiments in the present specification are described in a progressive manner, and the same and similar parts among the embodiments are referred to each other, and each embodiment focuses on the differences from other embodiments.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents, improvements and the like that fall within the spirit and principle of the present invention are intended to be included therein.
Claims (10)
1. An image self-adaptive segmentation method based on artificial intelligence is characterized by comprising the following steps:
step 1, acquiring a target gray image, counting gray values of all pixel points on the image, constructing a gray histogram, and calculating the probability of each gray level based on the gray histogram;
step 2, performing sliding window processing on the target gray level image, calculating the chaos of central pixel points in the sliding window according to the frequency of the gray levels corresponding to the pixel points in the sliding window and the variance of the gray levels, and further acquiring the chaos of all the pixel points on the target gray level image;
step 3, correcting the gray value of the corresponding pixel point by utilizing the chaos degree to obtain the gray value of the corrected pixel point, and establishing a corrected gray histogram, wherein the corrected histogram has obvious double-peak characteristics;
and 4, determining a segmentation threshold value based on the double-peak characteristics of the modified gray level histogram, and performing threshold segmentation on the target gray level image.
2. The image adaptive segmentation method based on artificial intelligence according to claim 1, wherein the step of extracting a defect region from the target gray image is further included before the construction of the gray histogram, and specifically includes:
setting a prior threshold and carrying out threshold segmentation on the target gray level image to obtain a defective area and a non-defective area.
3. The image adaptive segmentation method based on artificial intelligence of claim 2, wherein the size of the defect area and the set defect area threshold is judged, and if the defect area is smaller than the set defect area threshold, the threshold segmentation of steps 1-4 is performed on the target gray image; if the area of the defect area is larger than the set defect area threshold, the threshold segmentation is not carried out.
4. The image adaptive segmentation method based on artificial intelligence according to claim 1, wherein the obtaining method of the chaos degree of the pixel point specifically comprises:
wherein Q a Expressing the chaos of the pixel point a, the size of a sliding window taking the pixel point a as the center is n multiplied by n, p (X) expresses the probability when the gray level of the pixel point in the sliding window is X, X expresses the set of gray level series of each pixel point in the sliding window, n 2 -1 represents the number of pixels within the sliding window except the center pixel, S i Representing the gray level of the ith pixel point in the sliding window,
and expressing the average value of the gray levels of all the pixel points in the sliding window.
5. The artificial intelligence-based image adaptive segmentation method according to claim 2, wherein the step of extracting the defect region of the target gray image specifically comprises:
the part of the gray value of the pixel point on the target gray image, which is smaller than the prior threshold value, is a defect area, and the part of the gray value of the pixel point on the target gray image, which is larger than the prior threshold value, is a non-defect area.
6. An artificial intelligence based image adaptive segmentation system, characterized in that the system comprises:
the data acquisition module is used for acquiring a target gray image, counting the gray value of each pixel point on the image, constructing a gray histogram and calculating the probability of each gray level based on the gray histogram;
the data analysis module is used for performing sliding window processing on the target gray level image, calculating the chaos of central pixel points in the sliding window according to the frequency of the gray levels corresponding to the pixel points in the sliding window and the variance of the gray levels, and further acquiring the chaos of all the pixel points on the target gray level image;
the threshold segmentation module is used for correcting the gray value of the corresponding pixel point by utilizing the chaos degree to obtain the gray value of the corrected pixel point and establishing a corrected gray histogram, wherein the corrected histogram has obvious double-peak characteristics; and determining a segmentation threshold value based on the double-peak characteristics of the modified gray level histogram, and performing threshold segmentation on the target gray level image.
7. The system of claim 6, wherein the data acquisition module further comprises a defect region analysis unit configured to set a priori threshold and perform threshold segmentation on the target gray-scale image to obtain a defect region and a non-defect region.
8. The system according to claim 7, wherein the defect area analysis unit is further configured to determine the area of the defect area and a threshold of the defect area, and if the area of the defect area is smaller than the threshold of the defect area, the target grayscale image is input to the data analysis module for subsequent operations; and if the area of the defect area is larger than the set defect area threshold value, outputting a target gray level image.
9. The system according to claim 6, wherein the method for obtaining the degree of confusion of the pixel points comprises:
wherein Q is a Expressing the chaos of the pixel point a, the size of a sliding window taking the pixel point a as the center is n multiplied by n, p (X) expresses the probability when the gray level of the pixel point in the sliding window is X, X expresses the set of gray level series of each pixel point in the sliding window, n 2 -1 represents the number of pixels within the sliding window except the center pixel, S i Representing the gray level of the ith pixel point in the sliding window,
and expressing the average value of the gray levels of all the pixel points in the sliding window.
10. The artificial intelligence based image adaptive segmentation system of claim 6, wherein the step of extracting the defect region of the target gray image is specifically:
the part of the gray value of the pixel point on the target gray image, which is smaller than the prior threshold value, is a defect area, and the part of the gray value of the pixel point on the target gray image, which is larger than the prior threshold value, is a non-defect area.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210356859.0A CN114937055B (en) | 2022-03-31 | 2022-03-31 | Image self-adaptive segmentation method and system based on artificial intelligence |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210356859.0A CN114937055B (en) | 2022-03-31 | 2022-03-31 | Image self-adaptive segmentation method and system based on artificial intelligence |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114937055A true CN114937055A (en) | 2022-08-23 |
| CN114937055B CN114937055B (en) | 2024-05-03 |
Family
ID=82862447
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210356859.0A Active CN114937055B (en) | 2022-03-31 | 2022-03-31 | Image self-adaptive segmentation method and system based on artificial intelligence |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114937055B (en) |
Cited By (24)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115100203A (en) * | 2022-08-25 | 2022-09-23 | 山东振鹏建筑钢品科技有限公司 | Steel bar polishing and rust removing quality detection method |
| CN115100212A (en) * | 2022-08-29 | 2022-09-23 | 卡松科技股份有限公司 | Method for detecting pollution degree of lubricating oil |
| CN115100196A (en) * | 2022-08-24 | 2022-09-23 | 聊城市洛溪信息科技有限公司 | Method for evaluating derusting effect of stamping part based on image segmentation |
| CN115100198A (en) * | 2022-08-24 | 2022-09-23 | 江苏思锐装备科技有限公司 | Secondary oil removal intensity control method and system based on oil removal effect evaluation |
| CN115115624A (en) * | 2022-08-25 | 2022-09-27 | 江苏濠玥电子科技有限公司 | Rolling damage detection method for anti-corrosion coating of cable bridge |
| CN115229355A (en) * | 2022-09-22 | 2022-10-25 | 江苏双赢锻压机床有限公司 | Laser cutting method for high-precision stamping forging |
| CN115272315A (en) * | 2022-09-27 | 2022-11-01 | 南通致和祥智能装备有限公司 | Municipal garden area measurement method based on unmanned aerial vehicle aerial photography |
| CN115393354A (en) * | 2022-10-27 | 2022-11-25 | 江苏集宿智能装备有限公司 | Laser cutting defect identification method |
| CN115689948A (en) * | 2023-01-05 | 2023-02-03 | 济宁智诚物业管理有限公司 | Image enhancement method for detecting cracks of building water supply pipeline |
| CN115802056A (en) * | 2023-01-31 | 2023-03-14 | 南通凯沃智能装备有限公司 | User data compression storage method for mobile terminal |
| CN115984284A (en) * | 2023-03-21 | 2023-04-18 | 黄海造船有限公司 | X-ray hull weld image detection method for ship maintenance |
| CN116152133A (en) * | 2023-04-20 | 2023-05-23 | 山东超越纺织有限公司 | Blending defect detection method based on artificial intelligence |
| CN116912248A (en) * | 2023-09-13 | 2023-10-20 | 惠州市耀盈精密技术有限公司 | Irregular hardware surface defect detection method based on computer vision |
| CN116977230A (en) * | 2023-09-22 | 2023-10-31 | 济宁市质量计量检验检测研究院(济宁半导体及显示产品质量监督检验中心、济宁市纤维质量监测中心) | Scanning electron microscope image optimization enhancement method |
| CN117011303A (en) * | 2023-10-08 | 2023-11-07 | 泰安金冠宏油脂工业有限公司 | Oil production quality detection method based on machine vision |
| CN117152180A (en) * | 2023-10-31 | 2023-12-01 | 山东克莱蒙特新材料科技有限公司 | Mineral casting defect detection method based on artificial intelligence |
| CN117237350A (en) * | 2023-11-14 | 2023-12-15 | 青岛三合山精密铸造有限公司 | Real-time detection method for quality of steel castings |
| CN117372435A (en) * | 2023-12-08 | 2024-01-09 | 智联信通科技股份有限公司 | Connector pin detection method based on image characteristics |
| CN117409029A (en) * | 2023-12-14 | 2024-01-16 | 顺通信息技术科技(大连)有限公司 | Geothermal system environmental impact assessment method and system |
| CN117853488A (en) * | 2024-03-07 | 2024-04-09 | 咸阳黄河轮胎橡胶有限公司 | Tire wear degree detection method |
| CN117934350A (en) * | 2024-01-22 | 2024-04-26 | 北京富邦智达自动化科技有限公司 | Image processing method for intelligent inspection robot of chemical plant |
| CN118096761A (en) * | 2024-04-26 | 2024-05-28 | 宝鸡康盛精工精密制造有限公司 | Battery box production defect detection method based on image processing |
| CN118608523A (en) * | 2024-08-08 | 2024-09-06 | 俐玛光电科技(北京)有限公司 | Electronic component defect identification method |
| CN118608523B (en) * | 2024-08-08 | 2024-11-12 | 俐玛光电科技(北京)有限公司 | Electronic component defect identification method |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1540576A (en) * | 2003-04-21 | 2004-10-27 | ��ʽ���緺̩ | Method for looking up optinum threshold from image separation |
| CN101059870A (en) * | 2007-05-23 | 2007-10-24 | 华中科技大学 | Image cutting method based on attribute histogram |
| US20100329559A1 (en) * | 2009-06-29 | 2010-12-30 | Canon Kabushiki Kaisha | Image processing apparatus and control method thereof |
| CN106251351A (en) * | 2016-07-28 | 2016-12-21 | 国网湖南省电力公司 | A kind of transmission line forest fire monitoring threshold computational methods based on Cloud transform |
| CN107369159A (en) * | 2017-06-29 | 2017-11-21 | 大连理工大学 | Threshold segmentation method based on multifactor two-dimensional gray histogram |
| US20190392311A1 (en) * | 2018-06-21 | 2019-12-26 | Deep Force Ltd. | Method for quantizing a histogram of an image, method for training a neural network and neural network training system |
| CN113393395A (en) * | 2021-06-17 | 2021-09-14 | 西安应用光学研究所 | High-dynamic infrared image segmentation threshold self-adaptive calculation method |
-
2022
- 2022-03-31 CN CN202210356859.0A patent/CN114937055B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1540576A (en) * | 2003-04-21 | 2004-10-27 | ��ʽ���緺̩ | Method for looking up optinum threshold from image separation |
| CN101059870A (en) * | 2007-05-23 | 2007-10-24 | 华中科技大学 | Image cutting method based on attribute histogram |
| US20100329559A1 (en) * | 2009-06-29 | 2010-12-30 | Canon Kabushiki Kaisha | Image processing apparatus and control method thereof |
| CN106251351A (en) * | 2016-07-28 | 2016-12-21 | 国网湖南省电力公司 | A kind of transmission line forest fire monitoring threshold computational methods based on Cloud transform |
| CN107369159A (en) * | 2017-06-29 | 2017-11-21 | 大连理工大学 | Threshold segmentation method based on multifactor two-dimensional gray histogram |
| US20190392311A1 (en) * | 2018-06-21 | 2019-12-26 | Deep Force Ltd. | Method for quantizing a histogram of an image, method for training a neural network and neural network training system |
| CN113393395A (en) * | 2021-06-17 | 2021-09-14 | 西安应用光学研究所 | High-dynamic infrared image segmentation threshold self-adaptive calculation method |
Non-Patent Citations (1)
| Title |
|---|
| 吴成茂 等: "灰度级加权的直方图模糊熵阈值分割法", 西安邮电大学学报, no. 05, pages 14 - 19 * |
Cited By (36)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115100196A (en) * | 2022-08-24 | 2022-09-23 | 聊城市洛溪信息科技有限公司 | Method for evaluating derusting effect of stamping part based on image segmentation |
| CN115100198A (en) * | 2022-08-24 | 2022-09-23 | 江苏思锐装备科技有限公司 | Secondary oil removal intensity control method and system based on oil removal effect evaluation |
| CN115100196B (en) * | 2022-08-24 | 2022-11-18 | 聊城市洛溪信息科技有限公司 | Method for evaluating derusting effect of stamping part based on image segmentation |
| CN115115624A (en) * | 2022-08-25 | 2022-09-27 | 江苏濠玥电子科技有限公司 | Rolling damage detection method for anti-corrosion coating of cable bridge |
| CN115100203B (en) * | 2022-08-25 | 2022-11-18 | 山东振鹏建筑钢品科技有限公司 | Method for detecting quality of steel bar polishing and rust removal |
| CN115115624B (en) * | 2022-08-25 | 2022-11-22 | 江苏濠玥电子科技有限公司 | Rolling damage detection method for anti-corrosion coating of cable bridge |
| CN115100203A (en) * | 2022-08-25 | 2022-09-23 | 山东振鹏建筑钢品科技有限公司 | Steel bar polishing and rust removing quality detection method |
| CN115100212A (en) * | 2022-08-29 | 2022-09-23 | 卡松科技股份有限公司 | Method for detecting pollution degree of lubricating oil |
| CN115229355B (en) * | 2022-09-22 | 2022-12-13 | 江苏双赢锻压机床有限公司 | Laser welding method for high-precision stamping forging |
| CN115229355A (en) * | 2022-09-22 | 2022-10-25 | 江苏双赢锻压机床有限公司 | Laser cutting method for high-precision stamping forging |
| CN115272315A (en) * | 2022-09-27 | 2022-11-01 | 南通致和祥智能装备有限公司 | Municipal garden area measurement method based on unmanned aerial vehicle aerial photography |
| CN115393354A (en) * | 2022-10-27 | 2022-11-25 | 江苏集宿智能装备有限公司 | Laser cutting defect identification method |
| CN115689948A (en) * | 2023-01-05 | 2023-02-03 | 济宁智诚物业管理有限公司 | Image enhancement method for detecting cracks of building water supply pipeline |
| CN115802056A (en) * | 2023-01-31 | 2023-03-14 | 南通凯沃智能装备有限公司 | User data compression storage method for mobile terminal |
| CN115984284A (en) * | 2023-03-21 | 2023-04-18 | 黄海造船有限公司 | X-ray hull weld image detection method for ship maintenance |
| CN116152133A (en) * | 2023-04-20 | 2023-05-23 | 山东超越纺织有限公司 | Blending defect detection method based on artificial intelligence |
| CN116912248A (en) * | 2023-09-13 | 2023-10-20 | 惠州市耀盈精密技术有限公司 | Irregular hardware surface defect detection method based on computer vision |
| CN116912248B (en) * | 2023-09-13 | 2024-01-05 | 惠州市耀盈精密技术有限公司 | Irregular hardware surface defect detection method based on computer vision |
| CN116977230A (en) * | 2023-09-22 | 2023-10-31 | 济宁市质量计量检验检测研究院(济宁半导体及显示产品质量监督检验中心、济宁市纤维质量监测中心) | Scanning electron microscope image optimization enhancement method |
| CN116977230B (en) * | 2023-09-22 | 2024-01-02 | 济宁市质量计量检验检测研究院(济宁半导体及显示产品质量监督检验中心、济宁市纤维质量监测中心) | Scanning electron microscope image optimization enhancement method |
| CN117011303A (en) * | 2023-10-08 | 2023-11-07 | 泰安金冠宏油脂工业有限公司 | Oil production quality detection method based on machine vision |
| CN117011303B (en) * | 2023-10-08 | 2024-01-09 | 泰安金冠宏油脂工业有限公司 | Oil production quality detection method based on machine vision |
| CN117152180B (en) * | 2023-10-31 | 2024-01-26 | 山东克莱蒙特新材料科技有限公司 | Mineral casting defect detection method based on artificial intelligence |
| CN117152180A (en) * | 2023-10-31 | 2023-12-01 | 山东克莱蒙特新材料科技有限公司 | Mineral casting defect detection method based on artificial intelligence |
| CN117237350B (en) * | 2023-11-14 | 2024-01-26 | 青岛三合山精密铸造有限公司 | Real-time detection method for quality of steel castings |
| CN117237350A (en) * | 2023-11-14 | 2023-12-15 | 青岛三合山精密铸造有限公司 | Real-time detection method for quality of steel castings |
| CN117372435B (en) * | 2023-12-08 | 2024-02-06 | 智联信通科技股份有限公司 | Connector pin detection method based on image characteristics |
| CN117372435A (en) * | 2023-12-08 | 2024-01-09 | 智联信通科技股份有限公司 | Connector pin detection method based on image characteristics |
| CN117409029B (en) * | 2023-12-14 | 2024-02-13 | 顺通信息技术科技(大连)有限公司 | Geothermal system environmental impact assessment method and system |
| CN117409029A (en) * | 2023-12-14 | 2024-01-16 | 顺通信息技术科技(大连)有限公司 | Geothermal system environmental impact assessment method and system |
| CN117934350A (en) * | 2024-01-22 | 2024-04-26 | 北京富邦智达自动化科技有限公司 | Image processing method for intelligent inspection robot of chemical plant |
| CN117853488A (en) * | 2024-03-07 | 2024-04-09 | 咸阳黄河轮胎橡胶有限公司 | Tire wear degree detection method |
| CN118096761A (en) * | 2024-04-26 | 2024-05-28 | 宝鸡康盛精工精密制造有限公司 | Battery box production defect detection method based on image processing |
| CN118096761B (en) * | 2024-04-26 | 2024-07-23 | 宝鸡康盛精工精密制造有限公司 | Battery box production defect detection method based on image processing |
| CN118608523A (en) * | 2024-08-08 | 2024-09-06 | 俐玛光电科技(北京)有限公司 | Electronic component defect identification method |
| CN118608523B (en) * | 2024-08-08 | 2024-11-12 | 俐玛光电科技(北京)有限公司 | Electronic component defect identification method |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114937055B (en) | 2024-05-03 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN114937055B (en) | Image self-adaptive segmentation method and system based on artificial intelligence | |
| CN115170576B (en) | Aluminum pipe surface defect detection method based on machine vision | |
| CN115601364B (en) | Golden finger circuit board detection method based on image analysis | |
| CN111260616A (en) | Insulator crack detection method based on Canny operator two-dimensional threshold segmentation optimization | |
| CN111179243A (en) | Small-size chip crack detection method and system based on computer vision | |
| EP1269394B1 (en) | Improved method for image binarization | |
| CN104794502A (en) | Image processing and mode recognition technology-based rice blast spore microscopic image recognition method | |
| CN110766017B (en) | Mobile terminal text recognition method and system based on deep learning | |
| CN107490582B (en) | Assembly line workpiece detection system | |
| CN116485764B (en) | Structural surface defect identification method, system, terminal and medium | |
| CN114820625B (en) | Automobile top block defect detection method | |
| CN113177924A (en) | Industrial production line product flaw detection method | |
| US6993187B2 (en) | Method and system for object recognition using fractal maps | |
| CN110648330B (en) | Defect detection method for camera glass | |
| EP1605407A1 (en) | Image processing device | |
| CN111753794B (en) | Fruit quality classification method, device, electronic equipment and readable storage medium | |
| CN112308854A (en) | Automatic detection method and system for chip surface flaws and electronic equipment | |
| CN116883408B (en) | Integrating instrument shell defect detection method based on artificial intelligence | |
| CN111968082A (en) | Product packaging defect detection and identification method based on machine vision | |
| CN110060239B (en) | Defect detection method for bottle opening of bottle | |
| CN115131359A (en) | Method for detecting pitting defects on surface of metal workpiece | |
| CN113971681A (en) | Edge detection method for belt conveyor in complex environment | |
| CN112508913A (en) | Cable section edge detection method based on image detection | |
| JP2004045356A (en) | Surface defect detection method | |
| JP3906221B2 (en) | Image processing method and image processing apparatus |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20240412 Address after: 361000 room 404, No. 50, Chengyi North Street, phase III, software park, Xiamen, Fujian a078 Applicant after: Xiamen Hongyue Product Design Co.,Ltd. Country or region after: China Address before: 226000 No. 500, Linyang Road, Qidong Economic Development Zone, Nantong City, Jiangsu Province Applicant before: Jiangsu yijiesi Information Technology Co.,Ltd. Country or region before: China |
|
| TA01 | Transfer of patent application right | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |