CN105989583B - A kind of image defogging method - Google Patents

Abstract

The invention discloses an image defogging method, which uses a new transmittance estimation method instead of the soft matting method, and obtains the transmittance through the gain intervention of the minimum component of the three primary color channels of the foggy image, which can avoid halo and Under the premise of fast effect, quickly obtain the approximate dark channel image intensity, thereby improving the real-time performance of the algorithm, and use the quadtree subdivision method to estimate the atmospheric light value, so that the accuracy of the atmospheric light value is improved, and finally through the transmittance and atmospheric light Value multiple images are dehazed. The invention improves the real-time performance of defogging and the fidelity of defogging images.

Description

A method of image defogging

technical field

The invention belongs to the technical field of digital image processing and machine vision, and in particular relates to an image defogging method.

Background technique

With the development of science and technology and the improvement of people's living standards, more and more visual systems are used in real life, such as: monitoring systems, intelligent transportation systems, etc., and the quality of human life and needs are gradually improving. These application systems are closely related to human life. It is closely related and directly affects human life.

The reduction of air quality and the reduction of visibility have had a great impact on people's life and the development of industry. Due to the light being absorbed and scattered by the turbid medium, the pictures taken outdoors in bad weather will reduce the light intensity, resulting in a change in the light intensity received by the optical sensor, which will eventually lead to a decrease in color fidelity and serious color distortion and cast. shift, and the reduction of image contrast, the loss of image detail information, insufficient clarity, and the recognition of image information is greatly reduced. In addition, most automatic systems rely heavily on the definition of the input image, and degraded images will cause the automatic system to fail to work properly, affecting and limiting the work of highway visual monitoring, intelligent navigation, and remote sensing monitoring.

Therefore, the defogging processing of foggy images has broad application prospects, and can be applied to underwater shooting, aerial photography, outdoor monitoring, and even medical images. Images and videos after defogging processing are more valuable and beneficial to many image understandings. And computer vision applications (such as aerial images), image classification, image/video retrieval, remote sensing and video analysis and recognition, bring a lot of convenience to people's lives. In the field of image processing, image defogging started relatively late. At present, the image defogging methods proposed by domestic and foreign scholars are not particularly perfect, and need to be improved and improved urgently. In particular, how to improve the real-time performance and fidelity of defogging processing is the key to solving the defogging problem.

Contents of the invention

In order to solve the technical problems raised by the above-mentioned background technology, the present invention aims to provide an image defogging method, which adopts an improved calculation method of transmittance and atmospheric light value to improve the real-time performance of defogging and the accuracy of defogging images.

In order to realize above-mentioned technical purpose, technical scheme of the present invention is:

A kind of image defogging method, comprises the following steps:

(1) Calculate the transmittance of the foggy image:

t(x,y)=(1-I ^m (x,y))+gρ (1)

In formula (1), t(x, y) is the transmittance, (x, y) represents the pixel coordinates, ρ is the calibration factor, and g is the gain constant:

In formula (2 ⁾ , |I ^m | and |d| are the total number of pixels of Im and d respectively, corresponding to ^Im and d:

d(x,y)=I ^m (x,y)-I ^d (x,y) (4)

In formula (3), _Ic represents the color channel of the defogged image;

In formula (4), Ω(x,y) represents a square area centered on (x,y), and min represents the minimum value;

(2) Use the quadtree subdivision method to estimate the atmospheric light value of the foggy image. The specific steps are as follows:

(a) Divide the foggy image into several rectangular sub-blocks of equal size;

(b) calculate the average pixel value of each rectangular sub-block, retain the largest rectangular sub-block with an average pixel value, and record the average pixel value of this rectangular sub-block as the maximum pixel value _Amax ;

(c) If the maximum pixel value A _max is smaller than the preset pixel threshold a, and the size of the rectangular sub-block corresponding to the maximum pixel value A _max is greater than or equal to the preset minimum window size, return to step (a) and further divide the rectangle sub-block, otherwise enter step (d);

(d) Convert the image from RGB space to YCbCr space, and corresponding to the finally reserved rectangular sub-block, select the maximum value of the luminance component of the rectangular sub-block as the atmospheric light value A _c ;

(3) Dehaze the image according to the transmittance obtained in step (1) and the atmospheric light value obtained in step (2):

In formula (5), J represents the image after defogging processing, I represents the foggy image, t ₀ is the transmittance threshold, and max represents the maximum value.

Further, in step (1), the value range of the calibration factor ρ is [0.8,1].

Further, the value of the calibration factor ρ is 0.9.

Further, in step (3), the value range of the transmittance threshold t ₀ is [0.05, 0.15].

Further, the value of the transmittance threshold t ₀ is 0.1.

The beneficial effect brought by adopting the above-mentioned technical scheme:

Compared with the original dark channel algorithm, the present invention uses a new transmittance estimation method instead of the soft matting method, and obtains the transmittance through the gain intervention of the minimum component of the three primary color channels of the foggy image, which can avoid halo and Under the premise of fast effect, the approximate dark channel image intensity can be quickly obtained, thereby improving the real-time performance of the algorithm, and the quadtree subdivision method is used to estimate the atmospheric light value, which improves the accuracy of the atmospheric light value, thereby achieving good real-time performance and maintaining The effect of defogging the image with high degree of realism.

Description of drawings

Figure 1 is a basic flow chart of the present invention.

Fig. 2 is a flow chart of calculating the atmospheric light value in the present invention.

Detailed ways

The technical solutions of the present invention will be described in detail below in conjunction with the accompanying drawings.

As shown in Figure 1, an image defogging method includes the following steps:

Step 1. Calculate the transmittance of the foggy image:

t(x,y)=(1-I ^m (x,y))+gρ (1)

In formula (1), t(x, y) is the transmittance, (x, y) represents the pixel coordinates, ρ is the calibration factor, and the value of ρ is usually in [0.8,1]. In this embodiment, ρ= 0.9, g is the gain constant:

In formula (2 ⁾ , |I ^m | and |d| are the total number of pixels of Im and d respectively, corresponding to ^Im and d:

d(x,y)=I ^m (x,y)-I ^d (x,y) (4)

In formula (3), _Ic represents the color channel of the defogged image;

In formula (4), Ω(x, y) represents a square area centered on (x, y), and min represents the minimum value.

Step 2. Estimate the atmospheric light value of the foggy image using the quadtree subdivision method. The specific steps are shown in Figure 2:

(a) Divide the foggy image into several rectangular sub-blocks of equal size;

(b) calculate the average pixel value of each rectangular sub-block, retain the largest rectangular sub-block with an average pixel value, and record the average pixel value of this rectangular sub-block as the maximum pixel value _Amax ;

(c) If the maximum pixel value A _max is smaller than the preset pixel threshold a, and the size of the rectangular sub-block corresponding to the maximum pixel value A _max is greater than or equal to the preset minimum window size, return to step (a) and further divide the rectangle sub-block, otherwise enter step (d);

(d) Convert the image from the RGB space to the YCbCr space, and corresponding to the finally reserved rectangular sub-block, select the maximum value of the luminance component of the rectangular sub-block as the atmospheric light value _Ac .

Step 3. Dehaze the image according to the transmittance obtained in step 1 and the atmospheric light value obtained in step 2:

In formula (5), J represents the image after defogging processing, I is a foggy image, max represents the maximum value, _t0 is the transmittance threshold, and the value range of _t0 is [0.05,0.15]. In this embodiment You can set t ₀ =0.1 to limit the transmittance t(x,y), because when t(x,y)→0, J(x,y)t(x,y) will also approach 0, Make the dehazed image contain noise, in order to avoid this situation, set t ₀ to improve the quality of the dehazed image.

The above embodiments are only to illustrate the technical ideas of the present invention, and can not limit the protection scope of the present invention with this. All technical ideas proposed in accordance with the present invention, any changes made on the basis of technical solutions, all fall within the protection scope of the present invention. Inside.

Claims (5)

1. an image defogging method, is characterized in that, comprises the following steps: (1) Calculate the transmittance of the foggy image: t(x,y)=(1-I ^m (x,y))+gρ (1) In formula (1), t(x, y) is the transmittance, (x, y) represents the pixel coordinates, ρ is the calibration factor, and g is the gain constant: In formula (2 ⁾ , |I ^m | and |d| are the total number of pixels of Im and d respectively, corresponding to ^Im and d: d(x,y)=I ^m (x,y)-I ^d (x,y) (4) In formula (3), _Ic represents the color channel of the input image; In formula (4), Ω(x,y) represents a square area centered on (x,y), and min represents the minimum value; (2) Use the quadtree subdivision method to estimate the atmospheric light value of the foggy image. The specific steps are as follows: (a) Divide the foggy image into several rectangular sub-blocks of equal size; (b) calculate the average pixel value of each rectangular sub-block, retain the largest rectangular sub-block with an average pixel value, and record the average pixel value of this rectangular sub-block as the maximum pixel value _Amax ; (c) If the maximum pixel value A _max is smaller than the preset pixel threshold a, and the size of the rectangular sub-block corresponding to the maximum pixel value A _max is greater than or equal to the preset minimum window size, return to step (a) and further divide the rectangle sub-block, otherwise enter step (d); (d) Convert the image from RGB space to YCbCr space, and corresponding to the finally reserved rectangular sub-block, select the maximum value of the luminance component of the rectangular sub-block as the atmospheric light value A _c ; (3) Dehaze the image according to the transmittance obtained in step (1) and the atmospheric light value obtained in step (2): In formula (5), J represents the image after defogging processing, I represents the foggy image, t ₀ is the transmittance threshold, and max represents the maximum value. 2. An image defogging method according to claim 1, characterized in that: in step (1), the value range of the calibration factor ρ is [0.8, 1]. 3. An image defogging method according to claim 2, characterized in that: the value of the calibration factor ρ is 0.9. 4. A method for image defogging according to claim 1, characterized in that: in step (3), the value range of the transmittance threshold _t0 is [0.05, 0.15]. 5. An image defogging method according to claim 4, characterized in that: the value of the transmittance threshold _t0 is 0.1.