CN102788803B - On-line high-speed detection system of train wheel set tread and wheel rim and detection method thereof - Google Patents

Abstract

The invention discloses an on-line high-speed detection system and detection method for train wheel treads and rims. The detection system includes: a target detection device, a speed measurement device, a position judgment device, an image acquisition device and an image detection device arranged sequentially outside the track. The processing device and control device, the image acquisition device includes multiple sets of line structured light sources and a high-speed area array camera, the intersection line formed by the intersection of the light plane and the wheel set is located between the train sandblasting tube/braking device and the rail, the imaging of the area array camera The field of view is covered and slightly larger than the intersection area; the detection steps include ① using the canny edge detection method to determine whether the wheel set is defective; ② calculating the depth of the defect to obtain the size of the defect; ③ splicing the light bands in each image sequentially into an image. The advantages of the present invention are: avoiding the influence of train sandblasting pipe and brake device on the detection process; fast detection speed; this system is suitable for online detection of 20km/h~300km/train; high precision, and can realize defect detection below 1mm ; The detection results are presented in one image, which is convenient for observation.

Description

On-line high-speed detection system and detection method of train wheel tread and rim

technical field

The invention relates to a detection system and a detection method thereof, in particular to an on-line high-speed detection system and a detection method of a train wheel set tread and wheel rim.

Background technique

The wheel set is one of the most important parts of the train, and the quality of the wheel set directly affects the driving safety of the train. Defects on the tread surface of train wheels include tread scratches and tread peeling. These defects will bring additional impact and vibration during train operation, which will seriously affect the safety of trains and the service life of track facilities. So how to accurately detect train wheel sets? The defect of the tread surface is a detection technical problem that needs to be solved urgently in the train development.

In recent years, with the rapid development of my country's economy, the speed of trains has gradually increased. With the actual situation of high speed and heavy load, the train wheel sets have to bear greater dynamic loads, making the train wheels very prone to tread scratches, Phenomena such as peeling, which in turn lead to a decrease in the quality of train operation. The tread wear of the train wheels is intensified, and in severe cases, the axles will break, the wheels will collapse, and even major accidents will be caused.

At present, there are many detection methods for tread defects. Among them, in the patent No. ZL200610155281.3, the influence of the sandblasting tube and the brake device is not considered in the process of detecting the tread of the wheel set. The light plane of the structured light group is different from the track plane. The included angle and line array camera are used for imaging. Since the structured light group cannot completely cover the tread, it can only detect part of the defect information of the entire wheel set tread. In addition, the defect information is obtained by comparing with the non-defective picture and then doing the difference. , without considering the vibration and offset of the locomotive, it is easy to cause a large detection error.

Contents of the invention

In order to solve the deficiencies of the prior art, the object of the present invention is to provide a more accurate on-line detection system and detection method for detecting defects of train wheelsets.

In order to achieve the above object, the present invention adopts the following technical solutions:

The online high-speed detection system for train wheel tread and rim defects is characterized in that it includes:

A target detection device for detecting whether the train enters the detection area,

The speed measuring device used to detect the speed of the train,

A position judging device for judging whether the train has arrived at a designated position,

An image acquisition device for obtaining train wheel tread and rim image information,

an image processing device for processing the image information obtained by the aforementioned image acquisition device,

A system control device for receiving information from the aforementioned target detection device, speed measurement device and position judgment device, and controlling the image acquisition device and image processing device according to the obtained information;

The aforementioned image acquisition device includes: multiple groups of line-structured light sources and high-speed area array cameras; the light plane of the aforementioned line-structured light source is located above the track plane and forms a depression angle with the track plane, and the intersection line formed by the intersection of the light plane and the wheel set is parallel to the track plane and between the train sandblasting pipe/braking device and the track; the aforementioned high-speed area array camera is located below the track plane, and the optical axis of the camera forms an elevation angle with the track plane while forming a fixed angle with the aforementioned light plane; the aforementioned image acquisition device is At least two sets are respectively arranged on the outer sides of the two tracks.

The above-mentioned on-line high-speed detection system for tread and rim defects of train wheels is characterized in that the above-mentioned target detection device, position judgment device and speed measurement device are Hall sensors.

The aforementioned online high-speed detection system for train wheel tread and wheel rim defects is characterized in that the imaging field of view of the aforementioned high-speed area array camera can be set, and the width of the aforementioned imaging field of view only covers the area that reflects the shape of the wheel set tread and wheel rim , the height of the imaging field of view only covers the maximum intersection area.

The aforementioned online high-speed detection system for train wheel tread and rim defects is characterized in that the fixed angle formed by the optical axis of the aforementioned high-speed area array camera and the optical plane is within the range of 5 to 20 degrees.

The method for detecting wheel tread and rim defects by using the aforementioned train wheel tread and rim defect online high-speed detection system is characterized in that it includes the following steps:

(1) When the wheel set enters the detection area, the target detection device detects the corresponding information, and the system control device receives the aforementioned information and activates the speed measuring device;

(2) When the wheel set enters the speed measurement interval, the speed measurement device obtains the speed information of the wheel set, and the system control device receives the aforementioned speed information and starts the image acquisition device;

(3) When the wheel set reaches the designated image collection position, the position judging device detects the position information of the wheel set, the system control device receives the aforementioned position information, and controls the image acquisition device to collect images of the wheel set tread and rim, and obtains A series of tread and rim image information;

(4) The image processing device receives the aforementioned image information, processes and analyzes it, and finally displays the defect information of the tread and rim of the wheel set.

The aforementioned method for detecting defects on the tread and rim of the wheel set is characterized in that, in the aforementioned step (4), the image processing device processes the image information as follows:

(1) Judging whether there is a defect in the wheel set in the image: use the canny edge detection method to obtain the upper and lower single pixel edges of the intersection line, and set a first threshold at the same time to make a difference between the pixel point and the previous pixel point, when the difference is If it is greater than the first threshold, it is determined that there is a defect in the wheel set, otherwise, there is no defect;

(2) Calculate the size of the defect:

①. Take the intersection point of the tread surface and the line structure light source as the reference point of the single pixel edge, set a second threshold, and make a difference between several single pixel points before and after the estimated position separating the tread surface and the wheel rim from the reference point, and the difference is less than or equal to When the second threshold is reached, it is determined that this pixel point is the dividing point between the tread and the rim, and at the same time the actual number of pixels A of the rim is obtained;

②. Fit the points in the tread area to get a straight line. The deviation between the fitted point p(x,y) and the corresponding actual measurement point q(x,y) is r(x,y),r(x,y) =p(x,y)-q(x,y), the angle between the tangent plane at the position of the intersection line and the light plane is θ, then the defect depth d is: d=r(x,y)×tanθ;

③. According to the ratio between the rim and the length of the tread, calculate the theoretical pixel number B of the rim without wear, and make a difference between B and A to obtain the wear amount of the rim;

(3) Generate the result map:

The light band part in each image is extracted, spliced into an image in sequence, and the defect part of the wheel set is represented by a color different from the background.

The aforementioned method for detecting defects on the tread and rim of the wheel set is characterized in that the two pixel points that make the difference are separated by four pixel points.

The invention is beneficial in that: the intersection line formed by the intersection of the optical plane of the image acquisition device and the wheel set is located between the train sandblasting pipe/braking device and the rail, which solves the problem of detecting the train sandblasting pipe/braking device on the detection system The influence of direction and range; the intersection line formed by the intersection of the light plane of the line-structured light source and the wheel set is parallel to the track plane, and the transverse length of the line-structured light source covers the tread and rim of the wheel set, truly realizing the tread and rim All inspections without omission; the width of the imaging field of view of the high-speed area scan camera only covers the area that reflects the shape of the wheel tread and rim, and the height only covers the area of the largest intersection line, because no image is collected outside the intersection area Therefore, the calculation speed and work efficiency of the image processing device are greatly improved. This system is applicable to the online high-speed detection of 20km/h~300km/h trains; the method of the present invention is used to calculate the depth of tread defects without considering the vibration of the locomotive and offset, which greatly reduces the detection error; the light band part in each image is extracted and spliced into an image in sequence, and the defect part of the wheel set is also spliced in the aforementioned image at the same time, which is more convenient Observe the test results.

Description of drawings

Fig. 1 is the top view of the overall layout of the train wheelset defect online detection system of the present invention;

Fig. 2 is the right side view of the train wheelset defect online detection system in Fig. 1;

Fig. 3 is a schematic diagram of the relative position of the light plane of the line-structured light source and the tangent plane of the intersection line and the sandblasting tube/braking device;

The meanings of reference signs in the figure: 1-wheel set, 2-track, 3-target detection device, 4-position judgment device, 5-line structured light source, 6-high-speed area array camera, 7-speed measuring device, 8-light Plane, 9-tangent plane, 10-sandblasting tube, θ-angle, arrow indicates the direction of the train.

Detailed ways

The present invention will be specifically introduced below in conjunction with the accompanying drawings and specific embodiments.

With reference to Fig. 1, the online high-speed detection system of train wheelset defect of the present invention comprises: the target detection device 3 that is used to detect whether train enters the detection area, the velocity measuring device 7 that is used to detect train running speed, is used for judging whether train arrives designated position. Position determination device 4, an image acquisition device for obtaining image information of the tread and wheel rim of the train wheel pair 1, an image processing device (not shown) for processing the image information obtained by the image acquisition device, and a receiving target detection device 3. A system control device (not shown) that sends control commands to the image acquisition device based on the information from the speed measuring device 7 and the position judgment device 4 . The overall layout of the online detection system is detailed as follows:

On the outer side of the track 2 close to the track 2, arranged in sequence along the extending direction of the track 2: a target detection device 3, a speed measuring device 7, a position judging device 4, and an image acquisition device. The target detection device 3 first detects whether a train enters the detection area, and the speed measuring device 7 detects whether the running speed of the train after entering the detection area can satisfy the effective and complete image information collected by the image acquisition device. After specifying the image acquisition position, the position judging device 4 sends a signal to the system control device, and then starts the image acquisition device to collect images.

If the train runs faster, it is not conducive to the effective and complete image information collected by the image acquisition device, which will often cause the image processing done by the image processing device to be useless. Therefore, the speed measuring device 7 is connected with the system control device to establish a signal connection. When the detected train speed is less than a certain value, the system control device receives the information and sends an activation command to the image acquisition device to prepare for image acquisition; When the train speed exceeds this value, the speed measuring device 7 does not send a signal to the system control device, and the image acquisition device does not work, which avoids the useless work of the image acquisition device and the image processing device.

The image acquisition device includes: multiple groups of line-structured light sources 5 and a high-speed area array camera 6 forming an array. Referring to Fig. 2 and Fig. 3, the light plane 8 of the line-structured light source 5 is located above the track plane and forms a depression angle with the track plane, and the intersection line formed by the intersection of the light plane 8 and the wheel set 1 is parallel to the track plane and located on the train sandblasting tube. 10/Between the brake device and the track 2, the problem of the impact of the train sandblasting tube 10/brake device on the detection direction and range of the detection system is solved; the high-speed area array camera 6 is located below the track plane, and the optical axis of the camera is in line with the track plane An elevation angle is formed, and a fixed angle is formed with the light plane 8 at the same time.

As a preferred solution, the fixed angle formed by the optical axis of the high-speed area scan camera 6 and the optical plane 8 is in the range of 5-20 degrees. The larger the included angle, the more obvious the characteristics of the defects on the surface of the tread, but considering the number of installed line structure light sources 5, the included angle cannot be too large, so after testing, the size of the included angle is finally determined at 5~ within 20 degrees.

As a preferred solution, the position judging device 4 is in one-to-one correspondence with the image acquisition device, divided into two groups on average and arranged symmetrically on the outermost sides of the two tracks 2 . In order for the image acquisition device to obtain all the image information of the entire wheel set 1, that is, to collect all the image information of the wheel set 1 round, the number of the outermost image acquisition devices on each side of the track 2 is ≥ 2. A plurality of image acquisition devices are arranged at different positions but work at the same time, so that complete image information of the entire wheel set 1 can be obtained, which greatly increases the detection range.

As a preferred solution, the target detecting device 3, the position judging device 4 and the speed measuring device 7 are all Hall sensors. It is not only sensitive to magnetic fields, small in size, fast in response, but also has a long service life, can make accurate judgments, and transmit signals to the system control device, thereby ensuring that the online detection system of the present invention completes detection work with high quality.

As a preferred solution, the imaging field of view of the high-speed area array camera 6 can be set, the width of the imaging field of view only covers the area that reflects the shape of the wheel set tread and the rim, and the height of the imaging field of view only covers the maximum intersection area , which can be slightly larger than the intersection area. Since the imaging field of view of the high-speed area array camera 6 is not changed after being set, as the wheel set 1 approaches the image acquisition device, the proportion of the image formed by the intersection area in the entire image is also increasing. Therefore, the size of the imaging field of view of the camera should be able to cover or be slightly larger than the largest intersection area. Because most of the scenes outside the intersection area are not captured in the image, the image processing device only processes the image information in the intersection area when processing the image, which greatly improves the calculation speed and work efficiency. The driving speed is 20km/h~300km /h trains can use this system to detect wheel set defects.

The method for detecting wheel set defects using the above-mentioned online high-speed detection system for train wheel set tread and rim defects comprises the following steps:

First, after the wheel set 1 enters the detection area, the target detection device 3 detects the corresponding information, and sends a signal to the system control device, and the system control device activates the speed measuring device 7 . Then, after the wheel pair 1 enters the speed measuring interval, the speed measuring device 7 detects the speed of the wheel pair 1, if the speed is less than or equal to a certain value, the speed measuring device 7 transmits the speed signal to the system control device, and the system control device starts the image acquisition device to perform Good preparation for collecting image information; if the speed is greater than this value, in order to avoid that the image information collected by the image acquisition device is incomplete, that is, the image information of the entire wheel pair 1 cannot be obtained, the speed measuring device 7 does not send a signal to the system control device, that is, do not start the image acquisition device. After the image acquisition device is activated, when the wheel set 1 reaches the designated image acquisition position, the position judging device 4 detects the corresponding position information, and sends a signal to the system control device, and the system control device controls the image acquisition device to control the wheel set tread and Image acquisition is performed on the wheel rim to obtain a series of tread and wheel rim image information. Finally, the image processing device receives the image information, processes and analyzes it, and finally displays the defect information of the tread and rim of the wheel set.

Wherein, the image processing device processes the image information as follows:

First, determine whether the wheelset in the image is defective. The specific judgment process is as follows: use the canny edge detection method to obtain the upper and lower single pixel edges of the intersection line, which is the intersection line formed by the intersection of the light plane and the wheel set; set a first threshold, and perform deal with. As a preferred solution, each pixel is respectively compared with the fifth pixel in front of the pixel, and when the difference is greater than the first threshold, it is determined that there is a defect, otherwise, there is no defect.

Then, calculate the size of the defect. Take the intersection point of the tread and the line structure light source as the reference point of the single pixel edge, and set a second threshold, because the proportional relationship between the rim and the tread length is fixed, and an estimated position of the separation between the tread and the rim can be determined first through the total pixels. When the difference between several single pixel points before and after the estimated position and the reference point is less than or equal to the second threshold, it is determined that this pixel point is the boundary point for distinguishing the tread from the rim, and at the same time, the actual number of pixels A of the pixels occupied by the rim is calculated ;Fit the points in the tread area to get a straight line, the deviation between the tread fitting point p(x,y) and the corresponding actual measurement point q(x,y) is r(x,y),r(x, y)=p(x,y)-q(x,y), the angle between the tangent plane 9 at the position of the intersection line and the light plane 8 is θ, then the defect depth d is: d=r(x,y)× tanθ; According to a certain fixed ratio between the rim and the tread length, calculate the theoretical pixel number B of the rim without wear, and then make a difference between B and A. The difference is the pixel of the part of the rim that is reduced due to wear number, so as to obtain the amount of wear of the rim.

Finally, the image processing device stitches all the images together to form a detection result map. The specific splicing method is: extract the light band part in each image, splice into an image in sequence, and the defect part of the wheel set is also spliced in the above image. In order to facilitate observation, the defect part is used with another The background is represented by a different color.

It should be noted that the above-mentioned embodiments do not limit the present invention in any form, and all technical solutions obtained by means of equivalent replacement or equivalent transformation fall within the protection scope of the present invention.

Claims (6)

1. An online high-speed detection system for train wheel tread and rim defects, characterized in that it includes: A target detection device for detecting whether the train enters the detection area, The speed measuring device used to detect the speed of the train, A position judging device for judging whether the train has arrived at a designated position, An image acquisition device for obtaining train wheel tread and rim image information, an image processing device for processing the image information obtained by the above-mentioned image acquisition device, A system control device for receiving information from the above-mentioned target detection device, speed measurement device and position judgment device, and controlling the image acquisition device and image processing device according to the obtained information; The above-mentioned image acquisition device includes: multiple groups of line-structured light sources and high-speed area array cameras; the light plane of the above-mentioned line-structured light source is located above the track plane and forms a depression angle with the track plane, and the intersection line formed by the intersection of the light plane and the wheel set is parallel to the track plane and is located between the train sandblasting pipe/braking device and the track; the above-mentioned high-speed area array camera is located below the track plane, and the optical axis of the camera forms an elevation angle with the track plane and forms a fixed angle with the above-mentioned light plane at the same time; the above-mentioned image acquisition device At least two sets are arranged on the outer sides of the two tracks; the width of the imaging field of view of the above-mentioned high-speed area array camera only covers the area reflecting the wheel tread and the shape of the rim, and the height of the imaging field of view only covers the maximum intersection line area. 2. The online high-speed detection system for train wheel tread and rim defects according to claim 1, characterized in that the above-mentioned target detection device, position judgment device and speed measuring device are Hall sensors. 3. The online high-speed detection system for train wheel tread and rim defects according to claim 1, characterized in that the fixed angle formed by the optical axis of the high-speed area array camera and the optical plane is in the range of 5 to 20 degrees Inside. 4. Utilize the method for detecting wheel tread and rim defect of train wheel tread and rim defect on-line high-speed detection system according to claim 1, it is characterized in that, comprises the following steps: (1) When the wheel set enters the detection area, the target detection device detects the corresponding information, the system control device receives the above information, and starts the speed measurement device; (2) When the wheel set enters the speed measurement interval, the speed measurement device obtains the speed information of the wheel set, and the system control device receives the above speed information and starts the image acquisition device; (3) When the wheel set reaches the designated image collection position, the position judging device detects the position information of the wheel set, the system control device receives the above position information, and controls the image acquisition device to collect images of the wheel set tread and rim, and obtains A series of tread and rim image information; (4) The image processing device receives the above image information, processes and analyzes it, and finally displays the defect information of the wheel set tread and rim. 5. The method for detecting defects on the tread and rim of a wheel set according to claim 4, characterized in that, in the above step (4), the image processing device processes the image information as follows: (1) Judging whether there is a defect in the wheel set in the image: use the canny edge detection method to obtain the upper and lower single pixel edges of the intersection line, and set a first threshold at the same time to make a difference between the pixel point and the previous pixel point, when the difference is If it is greater than the first threshold, it is determined that there is a defect in the wheel set, otherwise, there is no defect; (2) Calculate the size of the defect: ①. Take the intersection point of the tread surface and the line structure light source as the reference point of the single pixel edge, set a second threshold, and make a difference between several single pixel points before and after the estimated position separating the tread surface and the wheel rim from the reference point, and the difference is less than or equal to When the second threshold is reached, it is determined that this pixel point is the dividing point between the tread and the rim, and at the same time the actual number of pixels A of the rim is obtained; ②. Fit the points in the tread area to get a straight line. The deviation between the fitting point p(x,y) and the corresponding actual measurement point q(x,y) is r(x,y), r(x,y) =p(x,y)-q(x,y), the angle between the tangent plane at the position of the intersection line and the light plane is θ, then the defect depth d is: d=r(x,y)×tanθ; ③. According to the ratio between the rim and the length of the tread, calculate the theoretical pixel number B of the rim without wear, and make a difference between B and A to obtain the wear amount of the rim; (3) Generate the result map: Extract the light band part from each image, splice it into an image in order, and express the defect part of the wheel set with a color different from the background. 6 . The method for detecting defects on the tread and rim of a wheel set according to claim 5 , wherein there are four pixels between the two pixels that make the difference.