An image preprocessing method for infrared target detection and tracking

Abstract

The present application relates to a kind of image pre-processing methods for infrared target detection tracking, belong to infrared target image pre-processing technical field, solve the technical problem of unstable infrared target detection tracking efficiency and tracking accuracy.It includes reading original infrared image I, carries out gray inversion;The probability density function pdf of the inverted infrared image I inv Is calculated, and the updated probability density function pdf new It is obtained by being handled to cut off;Image gray prior value p is calculated, and the probability density function pdf new It is divided into image high-order probability density function pdf h And image low-order probability density function pdf l Two parts;Image high-order probability density function pdf h And low-order probability density function pdf l Do weight calculation and accumulation calculation, do weighted adaptive gamma correction to inverted infrared image I inv ;To the corrected infrared image I e Do gray inversion;To infrared image I ed Carry out gray mapping, and the step of obtaining the pre-processed displayable image I ans . The method is used for the image pre-processing of infrared target detection tracking.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of infrared target image processing, and particularly relates to an image preprocessing method for infrared target detection and tracking. BACKGROUND

[0002] Target detection and tracking refers to inferring and predicting the position and scale of any object in a subsequent image frame in the initial state in a given first image frame, and is an indispensable and highly demanding work in the field of computer vision and pattern recognition, and has been widely applied in many fields. It is found in a large number of engineering application cases that infrared target detection and tracking based on optical imaging platforms such as airborne, vehicle-mounted and ship-mounted optical imaging platforms usually faces many difficulties such as low optical imaging signal-to-noise ratio, weak edge texture of the target, and serious background interference, so that the efficiency and tracking accuracy of infrared target detection and tracking are unstable.

[0003] In infrared target detection and tracking, the infrared target is usually located on the cloud layer or submerged in the background. In order to realize more stable target detection and tracking, the infrared image before detection and tracking needs to be preprocessed to remove a large amount of background noise to suppress background interference and only keep the target pixel information. Therefore, it is of extremely important significance to design an image preprocessing method for infrared target detection and tracking, realize background suppression of the infrared image, and highlight the infrared target pixel information for improving the stability and accuracy of infrared target detection and tracking in complex scenes. SUMMARY

[0004] In order to overcome the instability of the efficiency and tracking accuracy of infrared target detection and tracking, realize background suppression of the infrared image, and highlight the infrared target pixel information, the application provides an image preprocessing method for infrared target detection and tracking.

[0005] The technical solution adopted by the application to solve the technical problem is:

[0006] An image preprocessing method for infrared target detection and tracking, comprising the following steps:

[0007] Step 1: reading the original high-bit-depth infrared image I, performing gray scale inversion on the infrared image I to obtain an inverted infrared image I inv .

[0008] Step 2: calculating the probability density function pdf of the inverted image I inv , and performing truncation processing to limit the contrast, and uniformly distributing the frequency of the exceeding part to other pixel intensities to obtain the updated probability density function pdf new .

[0009] Step 3: calculate the image gray level priori value p, and divide the image probability density function pdf new into high-order probability density function pdf h and low-order probability density function pdf l ;

[0010] Step 4: respectively make weight calculation and accumulation calculation on the high-order probability density function pdf h and the low-order probability density function pdf l to obtain high-order cumulative distribution function cdf h and low-order cumulative distribution function cdf l , and make weighted adaptive gamma correction on the infrared image I inv based on the high-order cumulative distribution function cdf h and the low-order cumulative distribution function cdf l to obtain the corrected infrared image I e ;

[0011] Step 5: make gray level inversion on the corrected infrared image I e to obtain the infrared image I ed ;

[0012] Step 6: make gray level mapping on the infrared image I ed to obtain the preprocessed displayable image I ans .

[0013] Further, in Step 1, the infrared image I is inverted in gray level, and the calculation method is as follows:

[0014] I ive = l max -I

[0015] In the formula, l max represents the theoretical maximum gray level value of the image I, l max = 2 n -1, and n is the image bit depth.

[0016] Further, in Step 2, the probability density function pdf is truncated to limit the contrast, and the updated image probability density function pdf new is calculated as follows:

[0017] Step 1: calculate the truncation threshold N, and the calculation method is as follows:

[0018] N = clip limit × max (pdf)

[0019] In the formula, N represents the truncation threshold, cliplimit represents the set truncation percentage, and max represents the maximum value. Step 2: Establish an identification function id(·) to identify elements exceeding the truncation threshold, and the calculation method is as follows:

[0020]

[0021] Step 3: Residual calculation, calculate the sum N of elements exceeding the truncation threshold sum , and the calculation method is as follows:

[0022]

[0023] Step 4: Residual redistribution, evenly distribute the sum of elements exceeding the truncation threshold to other pixel intensities to obtain the updated image histogram pdf new , and the calculation method of the i-th gray level is as follows:

[0024]

[0025] Further, based on the image gray level prior value p, the updated probability density function pdf new is divided into high-order probability density function pdf h and low-order probability density function pdf l , and the calculation method is as follows:

[0026] Step 1: Calculate the image variance var, and the calculation method is as follows:

[0027]

[0028] In the formula, I ive (x,y) represents the gray value at pixel (x,y), and μ represents the gray mean value of the infrared image I ive .

[0029] Step 2: Calculate the image gradient prior val G , and the calculation method is as follows:

[0030]

[0031] In the formula, G x and G y are the gradients of the infrared image in the horizontal and vertical directions.

[0032] Step 3: Calculate the gray level prior value p, and the calculation method is as follows:

[0033]

[0034] In the formula, λ represents an adjustment parameter;

[0035] Further, the infrared image I inv Based on the high-order probability density function pdf h and the low-order probability density function pdf l of the image, weighted adaptive gamma correction is performed, and the calculation method is as follows:

[0036] Step 1: Calculate the high-order probability distribution function pdf w_h and the low-order probability density function pdf w_l of the weighted image, and the calculation method of the lth gray scale is as follows:

[0037]

[0038] In the formula, α and β respectively represent the adjustable power coefficients of the high-order probability distribution function pdf w_h and the low-order probability density function pdf w_l of the weighted image, and α>0, β>0;

[0039] Step 2: Calculate the cumulative distribution function cdf, and the calculation method of the lth gray scale is as follows:

[0040]

[0041] cdf(l) = cdf l (l) + cdf h (l)

[0042] In the formula, cdf h and cdf l respectively represent the high-order cumulative distribution function and the low-order cumulative distribution function of the image, and cdf w_h and cdf w_l are the cumulative distribution functions of the high-order probability distribution function pdf n and the low-order probability density function pdf n of the weighted image;

[0043] Step 3: Perform weighted adaptive gamma correction to obtain the updated gray scale T n , and the calculation formula of the lth gray scale is as follows:

[0044]

[0045] In the formula, T n (l) represents the pixel gray value corresponding to the pixel gray value l after the weighted gamma correction of the image cumulative distribution function cdf;

[0046] Step 4: Based on the updated gray scale T nThe original infrared image I is updated to obtain the corrected infrared image I e .

[0047] Further, the calculation method of the gray scale inversion in step 5 is the same as that in step 1.

[0048] Further, the pre-processed displayable image I ed is obtained by performing gray scale mapping on the infrared image I ans obtained in step 5 in step 6.

[0049]

[0050] MAX and MIN represent the maximum and minimum gray scales of the infrared image I ed , respectively, I ed (x, y) is the gray scale of the infrared image at the (x, y) position, and (x, y) is the pixel position of the image.

[0051] The present application has the following advantages:

[0052] An image pre-processing method for infrared target detection and tracking directly enhances the target and suppresses the background according to the spatial domain feature difference between the target and the background. The image is inverted in gray scale, and the segmented adaptive gamma correction based on the prior value is combined to effectively realize the strong enhancement of the bright target region of the image and the weak enhancement or suppression of the dark background region of the image. Then, the contrast of the inverted image is limited to effectively suppress the histogram component of the effective information played by the secondary information or even the noise in the image. The method suppresses the background noise enhancement, removes a large amount of background noise interference, highlights the infrared target pixel information, and can effectively improve the stability and precision of the infrared target detection and tracking in a complex scene. BRIEF DESCRIPTION OF DRAWINGS

[0053] Figure 1 The image pre-processing method flowchart of the present application embodiment 1 is shown in the figure.

[0054] Figure 2 The comparison chart of the background suppression and target enhancement effect of the image pre-processing method of the present application embodiment 1 on four infrared test images with strong noise interference and dark weak targets and four common methods is shown in the figure. DETAILED DESCRIPTION

[0055] The present application will be further described in detail below in combination with the drawings and specific embodiments.

[0056] Embodiment 1

[0057] An image preprocessing method for infrared target detection and tracking directly enhances the target and suppresses the background according to the spatial domain feature difference between the target and the background. Figure 1 As shown in the figure, first, the image is subjected to gray scale inversion, then the contrast of the inverted image is limited to suppress the background noise enhancement, and then the image histogram is divided into two parts by setting a gray scale prior value, the image is subjected to weighted adaptive gamma correction based on the two histogram parts, and finally the image is subjected to gray scale inversion and gray scale mapping to convert it into a preprocessed displayable image. Specifically, the following steps are included:

[0058] Step 1: read the original high-bit-depth infrared image I, perform gray scale inversion on the infrared image I to obtain the inverted infrared image I inv The calculation method of gray scale inversion is as follows:

[0059] I ive = l max -I

[0060] In the formula, l max represents the theoretical maximum gray scale value of the image I, l max = 2 n -1, and n is the image bit depth.

[0061] Step 2: calculate the probability density function pdf of the inverted image I inv , and perform truncation processing to limit the contrast, and uniformly distribute the frequency of the exceeding part to other pixel intensities to obtain the updated probability density function pdf new The calculation method is as follows:

[0062] First, calculate the truncation threshold, the calculation method is as follows:

[0063] N = cliplimit * max (pdf)

[0064] In the formula, N represents the truncation threshold, cliplimit represents the set truncation percentage, and max represents the maximum value.

[0065] Secondly, the identification function id(·) is established to identify the elements exceeding the truncation threshold, the calculation method is as follows:

[0066]

[0067] Then, the remainder is calculated, and the sum N sum exceeding the truncation threshold is calculated, the calculation method is as follows:

[0068]

[0069] Finally, the residual is redistributed, and the sum exceeding the truncation threshold is evenly distributed to other pixel intensities to obtain the updated image histogram pdf new The i-th gray level is calculated as follows:

[0070]

[0071] Step 3: Calculate the image gray level prior p, and based on the image gray level prior p, the image probability density function pdf new is divided into an image high-order probability density function pdf h and a low-order probability density function pdf l Two parts, the calculation method is as follows:

[0072] First, calculate the image variance var, the calculation method is as follows:

[0073]

[0074] In the formula, I ive (x, y) represents the gray value at the pixel (x, y), and μ represents the gray mean value of the inverted infrared image I ive ;

[0075] Secondly, calculate the image gradient prior val G , the calculation method is as follows:

[0076]

[0077] In the formula, G x and G y are the gradients of the infrared image in the horizontal direction and the vertical direction;

[0078] Finally, based on the image variance var and the image gradient prior val G , the gray level prior p is calculated, and the calculation method is as follows:

[0079]

[0080] In the formula, λ represents an adjustment parameter;

[0081] Step 4: Respectively, the image high-order probability density function pdf h and the low-order probability density function pdf l are calculated, and the image high-order cumulative distribution function cdf h and the low-order cumulative distribution function cdf l are obtained, and the infrared image I inv is based on the image high-order cumulative distribution function cdf h and the low-order cumulative distribution function cdf lThe weighted adaptive gamma correction is performed to obtain the corrected infrared image I e ;

[0082] First, the high-order probability distribution function pdf of the weighted image is calculated w_h and the low-order probability density function pdf w_l The calculation method of the lth gray scale is as follows:

[0083]

[0084] In the formula, α and β respectively represent the adjustable power coefficients of the high-order probability distribution function pdf of the weighted image w_h and the low-order probability density function pdf w_l , α>0, β>0;

[0085] Second, the image cumulative distribution function cdf is calculated, and the calculation method of the lth gray scale is as follows:

[0086]

[0087] cdf(l) = cdf l (l) + cdf h (l)

[0088] In the formula, cdf h and cdf l respectively represent the high-order cumulative distribution function and the low-order cumulative distribution function, and the cumulative distribution function of the high-order probability distribution function pdf of the weighted image w_h and the low-order probability density function pdf w_l ;

[0089] Then, the weighted adaptive gamma correction is performed based on the cumulative distribution function cdf to obtain the updated gray scale T n , and the calculation method of the lth gray scale value is as follows:

[0090]

[0091] In the formula, T n (l) represents the pixel gray scale value corresponding to the pixel gray scale value l after the weighted gamma correction based on the histogram pdf h and pdf l ;

[0092] Finally, the original infrared image I is updated based on the updated gray scale T n to obtain the corrected infrared image I e .

[0093] Step 5: The corrected infrared image I eInfrared image I is obtained by performing grayscale inversion. ed The calculation method for grayscale inversion is the same as in step 1;

[0094] Step 6: Process the infrared image I ed Grayscale mapping is performed to obtain a preprocessed displayable image I. ans The grayscale value at position (x,y) is calculated as follows:

[0095]

[0096] In the formula, MAX and MIN represent the infrared image I, respectively. ed The maximum and minimum grayscale values, I ed (x,y) represents the grayscale value of the infrared image at position (x,y), and (x,y) represents the position of the image pixel.

[0097] like Figure 2 The image shows the effect of applying this method to four infrared images with strong noise interference and faint targets. Compared with several common methods (HE, CLAHE, BBHE and AGCWD), it can be seen that the noise background is suppressed. At the same time, the contrast of the target is enhanced without introducing artifacts or noise, resulting in better visualization. This can provide high-quality target scenes for subsequent target tracking methods to achieve more stable and accurate target tracking.

Claims

1. An image preprocessing method for infrared target detection and tracking, characterized in that, The method comprises the following steps: Step 1, reading an original infrared image I, carrying out gray scale inversion on the read original infrared image I to obtain an inverted infrared image I inv ; Step 2, calculate the probability density function pdf of the reversed infrared image I inv , and truncate it, distributing the exceeding part evenly to other gray scale intensities to obtain the updated probability density function pdf new ; the calculation method is as follows: First, the truncation threshold is calculated as follows: N = clip limit * max (pdf) In the formula, N represents the truncation threshold, clip limit represents the set truncation percentage, and max represents the maximum value; second, the identification function id(·) is established to identify the elements exceeding the truncation threshold, and the calculation is as follows: Then a residual calculation is performed, calculating the sum N of the values that exceed the cutoff threshold sum in the following manner: Finally, the residual is redistributed, and the sum exceeding the clipping threshold is evenly distributed to other pixel intensities, resulting in an updated probability density function pdf new The i-th gray scale is calculated as follows: Step 3, calculating image gray level prior value p based on image variance var and image gradient prior val G ; dividing probability density function pdf new into image high-order probability density function pdf h and image low-order probability density function pdf l based on image gray level prior value p; Step 4, weight calculation and accumulation calculation are performed on the image high-order probability density function pdf h and the image low-order probability density function pdf l , to obtain the image high-order cumulative distribution function cdf h and the image low-order cumulative distribution function cdf l ; based on the image high-order cumulative distribution function cdf h and the image low-order cumulative distribution function cdf l , weighted adaptive gamma correction is performed on the inverted infrared image I inv , to obtain the corrected infrared image I e ; Step 5. Invert the corrected infrared image I e to obtain the infrared image I ed ; Step 6. Grayscale mapping is performed on the infrared image I ed to obtain a pre-processed displayable image I ans .

2. The image pre-processing method of claim 1, wherein, In step 1, the original infrared image I read is subjected to a gray scale inversion, and the inverted infrared image I inv The calculation formula is as follows: I inv = l max -I In the above formula, l max is the theoretical maximum gray value of the original infrared image I, l max = 2 n -1, n is the image bit depth.

3. The image preprocessing method of claim 1, wherein, The step 3 further comprises: The image gray level prior value p is calculated as follows: In the above equation, λ is a tuning parameter, val G is the image gradient prior, and μ is the mean of the image grayscale.

4. The image pre-processing method of claim 3, wherein, The image gray level prior value p is calculated, first, the image variance var is calculated as follows: In the above formula, I ive (x,y) is the gray value at pixel (x,y); Second, the image gradient prior val is calculated G The calculation formula is as follows: In the above formula, G x and G y are gradients of the infrared image in the horizontal direction and the vertical direction.

5. The image pre-processing method of claim 1, wherein, The step 4 further comprises: Based on the cumulative distribution function cdf(l), the weighted adaptive gamma correction is performed to obtain the updated gray scale T n The calculation formula of the first gray scale is as follows: In the above formula, T n (l) is the pixel gray value corresponding to the weighted gamma-corrected pixel gray value of the pixel gray value l; The updated gray scale T is obtained according to the weighted adaptive gamma correction n The original infrared image I is updated in gray scale, receiving a corrected infrared image I e .

6. The image pre-processing method of claim 1, wherein, The step 6 further comprises: The gray level I of the displayable image at the (x, y) position after pre-processing ans (x, y) is calculated as follows: In the above formula, MAX and MIN are the maximum and minimum gray scale values of the infrared image I ed ed (x, y) is the gray scale of the infrared image at the position (x, y), and (x, y) is the position of the image pixel.​

Images