Double-histogram equalization method based on image partitioning

Abstract

The invention provides a double-histogram equalization method based on image partitioning. The method comprises the following steps: S1, image partitioning; s2, performing operation on each sub-block of the image block; in each sub-block, S2.1, performing statistics on a histogram, and calculating a histogram clipping threshold value and the average gray scale of the block; s2.2, dividing the histogram into two sub-histograms according to the average gray scale; s2.3, respectively cutting the two histograms according to a cutting threshold value; s2.4, the cut-out parts in the two sub-histograms are distributed according to a preset distribution mode; s2.5, respectively superposing the sub-histograms on the residual sub-histograms; s2.6, respectively carrying out histogram equalization on the two sub-histograms in respective gray scale ranges, and combining the two obtained mapping functions as a final mapping function; and S3, performing bilinear interpolation operation on the image according to the pixel position and the mapping function of each block to obtain an enhanced image. The problems of over-enhancement and low local contrast of a histogram equalization method are solved, and it is ensured that a low-contrast image has a proper contrast after enhancement.

Description

Technical Field

[0001] The present invention belongs to the technical field of image processing, and particularly relates to a dual histogram equalization method based on image block division. Background Art

[0002] In the prior art, contrast enhancement is of great significance for improving the visual perception of digital images. An image with appropriate contrast is not only more natural visually, but also can retain more image information, facilitating subsequent image processing. However, during the image imaging process, after a series of processes such as denoising, gamma correction, and color space conversion of the raw data input by the sensor, the resulting output image has a low contrast. Therefore, a contrast enhancement method is needed that can enhance a low-contrast image into an image with a higher contrast while maintaining a more natural visual effect for the enhanced image. The traditional histogram equalization (HE) algorithm is the most widely used contrast enhancement method, but it often has the problem of over-enhancement. Moreover, since the enhancement process of HE operates on the entire current image, it often ignores local contrast, resulting in a low local contrast of the enhanced image.

[0003] Histogram equalization is a method that can effectively improve the contrast of an image. It calculates the cumulative distribution function (CDF), also called the mapping function, using the gray-scale distribution of the current image, and uses this function to map the gray-scale of the input image to achieve the purpose of enhancing the image contrast. However, when calculating the mapping function, this method only considers the histogram of the current image. When the gray-scale distribution of the current image is relatively concentrated, there will be dominant gray-scales, which affect the slope of the mapping function, resulting in over-enhancement of the input image. That is to say, a main problem of HE is that it causes over-enhancement of the image, that is, the contrast of the enhanced image is too high, affecting the visual effect of the image.

[0004] In addition, common terms in the prior art include:

[0005] ISP: image signal processer, a module that performs algorithm processing on the raw data input by the sensor;

[0006] HE: histogram equalization, a method for modifying image contrast; PDF: Probability Density Function;

[0007] CDF: Cumulative Distribution Function;

[0008] Dominant gray level: The gray level with a relatively large number of pixels in the image, which will affect the calculation of the CDF in histogram equalization;

[0009] Over-enhancement: The contrast of the enhanced image is too high, resulting in abnormal image effects. Summary of the Invention

[0010] To solve the above problems, the purpose of this application is to: solve the problems of over-enhancement and low local contrast in the histogram equalization method, and ensure that the enhanced low-contrast image has an appropriate contrast. Specifically, in order to ensure that the enhanced image has a higher local contrast, this method introduces the operation of image block division; at the same time, to solve the over-enhancement problem in the HE algorithm, a method of histogram segmentation, clipping, and redistribution is proposed; compared with the original HE algorithm, the enhanced image not only has better local contrast but also has a more natural visual effect.

[0011] Specifically, the present invention provides a dual histogram equalization method based on image block division, and the method includes the following steps:

[0012] S1, Image block division;

[0013] S2, Operations on each sub-block of the image block division; that is, within each sub-block:

[0014] S2.1, Statistic the histogram, calculate the histogram clipping threshold and the average gray level of this block;

[0015] S2.2, Divide the histogram into two sub-histograms according to the average gray level;

[0016] S2.3, Clip the two histograms respectively according to the clipping threshold;

[0017] S2.4, Allocate the clipped parts in the two sub-histograms according to a preset allocation method;

[0018] S2.5, And superimpose them on the remaining sub-histograms respectively;

[0019] S2.6, Perform histogram equalization on the two sub-histograms within their respective gray level ranges, and combine the two obtained mapping curves as the final mapping function;

[0020] S3, Perform bilinear interpolation operation on the image according to the pixel position and the mapping function of each block to obtain the enhanced image.

[0021] The step S1 is to divide the input luminance image into blocks, which further includes:

[0022] (1) Input the luminance image y, and the width and height of the input image are imgWidth and imgHeight respectively;

[0023] (2) Set the number of vertical blocks bloVer and the number of horizontal blocks bloHor, where bloVer = 8 and bloHor = 8; if the block division method is vertical block division, then:

[0024] sideVer = ((imgHeight / 2) / bloVer) × 2 (1)

[0025] remVer = (imgHeight / 2) % bloVer (2)

[0026] Among them, sideVer represents the minimum number of pixels for vertical block division, and remVer determines whether the pixel height of the image can be evenly divided by the number of vertical blocks during vertical block division; when remVer is not equal to 0, that is, there is a non - divisible situation during the block division process, an operation of +2 will be performed on the pixel height of the first remVer blocks of the vertical block division, as shown in the following formula (3), to ensure that this allocation method can satisfy the following formula (4); that is

[0027]

[0028] (sideVer + 2) × remVer + sideVer × (bloVer - remVer) = imgHeidth (4)

[0029] In formula (3), bloHeight(h) represents the pixel height of the h - th block during vertical block division, h ∈ [1, bloVer]; similarly, during horizontal block division:

[0030] sideHor = ((imgWidth / 2) / bloHor) × 2 (5)

[0031] remHor = (imgWidth / 2) % bloHor (6)

[0032]

[0033] (sideHor + 2) × remHor + sideHor × (bloHor - remHor) = imgWidth (8)

[0034] In formulas (5) - (7), sideHor represents the minimum number of pixels for horizontal block division, remHor determines whether the pixel width of the image can be evenly divided by the number of horizontal blocks during horizontal block division, and bloWidth(w) represents the pixel width of the w - th block during horizontal block division, w ∈ [1, bloHor];

[0035] (3) Divide the input image y into bloVer × bloHor blocks according to the calculated results.

[0036] Step S2 further includes:

[0037] S2.1, within each sub - block y h,w :

[0038] (1) Count the histogram information hist h,w of the current block y h,w , with a total of numBins gray levels;

[0039] (2) Calculate the average gray value avg h,w of y h,w , and the formula is as follows:

[0040]

[0041] where represents rounding down, hist h,w (i) represents the number of pixels with pixel value i in y h,w , and i ∈ [0, numBins);

[0042] (3) Calculate the clipping threshold clipThre h,w of this block according to the preset clipping ratio clipRatio, and the calculation process is as follows:

[0043]

[0044] Here, clipRatio = 3;

[0045] S2.2, within each sub - block y h,w , divide the histogram into two sub - histograms, a lower - level sub - histogram and a higher - level sub - histogram, according to the average gray value avg h,w ;

[0046] (1) Sub - histogram Figure 1 comes from the lower - gray - level area of hist h,w , denoted as histL h,w ; Sub - histogram 2 comes from the higher - gray - level area of hist h,w , denoted as histH h,w , and the specific division method is:

[0047] histL h，w (m) = hist h，w (i), 0 ≤ m, i < avg h，w (11)

[0048] histH h，w (n) = hist h，w (i), avg h，w≤i < numBins, 0 ≤ n < numBins - avg h，w (12)

[0049] where m and n represent the m-th and n-th gray levels of the sub-histograms histL h,w , histH h,w respectively, m ∈ [0, avg h，w ), n ∈ [0, numBins - avg h，w );

[0050] (2) The two sub-histograms need to satisfy:

[0051]

[0052] S2.3. Within each block, use clipThre obtained in (3) of step S2.1 h,w to clip histL h,w and histH h,w respectively,

[0053] (1) Denote the remaining parts as histRemL h,w and histRemH h,w , and the clipping process is as follows:

[0054]

[0055]

[0056] (2) Count the number of pixels numL and numH clipped at each gray level, and accumulate them respectively to obtain the total number of pixels in the clipped part, denoted as totalL h,w and totalH h,w :

[0057]

[0058]

[0059]

[0060]

[0061] (3) This clipping process needs to satisfy:

[0062]

[0063]

[0064] S2.4. Within each block, re - distribute the cropped parts of the two sub - histograms according to a preset distribution method, and denote the two obtained sub - histograms as allocL h，w and allocH h，w :

[0065]

[0066]

[0067] S2.5. Within each block, superimpose the re - distributed allocL h，w and allocH h，w onto the remaining parts histRemL h，w and histRemH h，w to obtain new sub - histograms histL h，w ' and histH h，w ':

[0068] histL h，w '(m)=histRemL h，w (m)+allocL h，w (m), 0 ≤ m < avg h，w

[0069] histH h，w '(n)=histRemH h，w (n)+allocH h，w (n), 0 ≤ n < numBins - avg h，w (19)

[0070] S2.6. Within each block, perform histogram equalization on the histograms histL h,w ' and histH h，w ' within their respective gray - scale ranges, including:

[0071] (1) Calculate the probability density function pdfL h，w ' of the gray - scale levels of histL h，w , the formula is as follows:

[0072]

[0073] (2) Calculate the probability density function pdfH h，w ' of the gray - scale levels of histH h，w , the formula is as follows:

[0074]

[0075] (3) The pdfL h，wThe probability density function is accumulated, and within its gray level range [0avg h，w -1], the mapping function cdfL h，w is calculated, and the formula is as follows:

[0076]

[0077] (4) The probability density function of pdfH h，w is accumulated, and within its gray level range [avg h，w numBins - 1], the mapping function cdfH h，w is calculated, and the formula is as follows:

[0078]

[0079] (5) cdfL h,w and cdfH h,w are combined to obtain the final mapping function cdf h,w ;

[0080]

[0081] The scope of action of step S2 is in each image block, and the function in step S2 is to calculate a mapping function within each image block by the method of histogram piecewise reassignment.

[0082] In step S2.1, numBins = 256.

[0083] In step S3, according to the pixel position and the mapping function cdf h,w in each block, a bilinear interpolation operation of the image is performed to obtain the enhanced image y'.

[0084] Thus, the advantages of this application are as follows:

[0085] 1. The present invention designs a double histogram equalization method based on image block division. This method not only solves the over-enhancement problem existing in the traditional histogram equalization algorithm, but also the enhanced image has better local contrast;

[0086] 2. The block division operation introduced by the present invention solves the problem that the local contrast cannot be maintained during the histogram equalization process;

[0087] 3. The image block calculation method proposed by the present invention can adaptively calculate the pixel height and pixel width of the block according to the input resolution and the number of blocks, and solves the problem that the input resolution cannot be divided evenly during the image block division process;

[0088] 4. The present invention proposes a method of dividing the block histogram into two sub-histograms, namely a high sub-histogram and a low sub-histogram, which solves the over-enhancement problem in the process of histogram equalization;

[0089] 5. The sub-histogram cropping and reallocation method proposed by the present invention ensures good local contrast of the input image after enhancement. BRIEF DESCRIPTION OF THE DRAWINGS

[0090] The drawings described herein are used to provide a further understanding of the present invention, and constitute a part of this application, but do not limit the present invention.

[0091] Figure 1 It is a schematic diagram of the process of this method. DETAILED DESCRIPTION OF THE INVENTION

[0092] In order to more clearly understand the technical content and advantages of the present invention, the present invention will be further described in detail below with reference to the drawings.

[0093] The present application proposes a dual-histogram equalization method based on image block division, which mainly includes: S1, image block division;

[0094] S2, within each sub-block:

[0095] S2.1, statistically calculate the histogram, calculate the histogram cropping threshold and the average gray level of this block;

[0096] S2.2, divide the histogram into two sub-histograms according to the average gray level;

[0097] S2.3, respectively crop the two histograms according to the cropping threshold;

[0098] S2.4, allocate the cropped parts in the two sub-histograms according to a preset allocation method;

[0099] S2.5, and respectively superimpose them on the remaining sub-histograms;

[0100] S2.6, perform histogram equalization on the two sub-histograms within their respective gray-level ranges, and combine the two obtained mapping curves as the final mapping function;

[0101] S3, perform bilinear interpolation operation on the image according to the pixel position and the mapping function of each block to obtain the enhanced image.

[0102] Specifically, as Figure 1 shown, the present invention proposes a dual-histogram equalization method based on image block division, which is used to solve the over-enhancement problem existing in the traditional histogram equalization algorithm, so that the enhanced image has good contrast. It includes:

[0103] S1, divide the input luminance image into blocks, which mainly includes:

[0104] (1) Input the luminance image y. The width and height of the input image are imgWidth and imgHeight respectively. Here, imgWidth = 1920 and imgHeight = 1080.

[0105] (2) Set the number of vertical blocks bloVer and the number of horizontal blocks bloHor. Here, bloVer = 8 and bloHor = 8. The block division method is (taking the vertical block division as an example first):

[0106] sideVer = ((imgHeight / 2) / bloVer) × 2 (1)

[0107] remVer = (imgHeight / 2) % bloVer (2)

[0108] Among them, sideVer represents the minimum number of pixels in a vertical block, and remVer determines whether the pixel height of the image can be evenly divided by the number of vertical blocks during vertical block division. When remVer is not equal to 0, that is, when there is a non - divisible situation during the block division process, an operation of +2 will be performed on the pixel height of the first remVer blocks in the vertical block division (as shown in the following formula (3)) to ensure that this allocation method can satisfy the following formula (4); that is

[0109]

[0110] (sideVer + 2) × remVer + sideVer × (bloVer - remVer) = imgHeight (4)

[0111] In formula (3), bloHeight(h) represents the pixel height of the h - th block during vertical block division, h ∈ [1, bloVer]. Here, sideVer = 135, remVer = 0, and correspondingly, bloHeight(1 - 8) = 135.

[0112] Similarly, during the horizontal block division process:

[0113] sideHor = ((imgWidth / 2) / bloHor) × 2 (5)

[0114] remHor = (imgWidth / 2) % bloHor (6)

[0115]

[0116] (sideHor + 2) × remHor + sideHor × (bloHor - remHor) = imgWidth (8)

[0117] In formulas (5)-(7), sideHor represents the minimum number of pixels for horizontal partitioning, remHor determines whether the pixel width of the image can be evenly divided by the number of horizontal partitions when performing horizontal partitioning, and bloWidth(w) represents the pixel width of the w-th block during horizontal partitioning, where w ∈ [1, bloHor]; here, sideHor = 240, remVer = 0, and correspondingly, bloWidth(1 to 8) = 240.

[0118] (3) Divide the input image y into a total of 64 blocks of bloVer × bloHor according to the calculated results.

[0119] S2. Perform operations on each image block:

[0120] S2.1. Within each block y h,w :

[0121] (1) Count the histogram information hist h,w of the current block y h,w , with a total of numBins gray levels; here, numBins = 256;

[0122] (2) Calculate the average gray level avg h,w of y h,w , and the formula is as follows:

[0123]

[0124] where represents rounding down, hist h,w (i) represents the number of pixels with pixel value i in y h,w , where i ∈ [0, numBins);

[0125] (3) Calculate the clipping threshold clipThre h,w of this block according to the preset clipping ratio clipRatio, and the calculation process is as follows:

[0126]

[0127] Here, clipRatio = 3;

[0128] S2.2. Within each block y h,w , divide the histogram into two sub-histograms, a high sub-histogram and a low sub-histogram, according to the average gray level avg h,w .

[0129] (1) The sub-histogram Figure 1 is derived from the lower gray level region of hist h,w , denoted as histL h,w; The sub-histogram 2 is derived from hist h,w from the higher gray-level region, denoted as histH h,w , and the specific division method is as follows:

[0130] hisL h，w (m) = hist h，w (i), 0 ≤ m, i < avg h，w (11)

[0131] hisH h，w (n) = hist h，w (i), avg h，w ≤ i < numBins, 0 ≤ n < numBins - avg h，w (12)

[0132] where m and n respectively represent the m-th and n-th gray levels of the sub-histograms histL h,w , histH h,w , m ∈ [0, avg h，w )), n ∈ [0, numBins - avg h，w )).

[0133] (2) The two sub-histograms need to satisfy:

[0134]

[0135] S2.3. Within each block, use clipThre obtained in (3) of step S2.1 h,w to clip histL h,w and histH h,w respectively,

[0136] (1) Denote the remaining parts as histRemL h,w and histRemH h,w , and the clipping process is as follows:

[0137]

[0138]

[0139] (2) Count the number of pixels clipped at each gray level, numL and numH, and accumulate them respectively to obtain the total number of pixels in the clipped part, denoted as totalL h,w and totalH h,w :

[0140]

[0141]

[0142]

[0143]

[0144] (3) This cropping process needs to satisfy:

[0145]

[0146]

[0147] S2.4. Within each block, re - distribute the cropped parts of the two sub - histograms according to a preset distribution method, and denote the two obtained sub - histograms as allocL h,w and allocH h,w :

[0148]

[0149]

[0150] S2.5. Within each block, add the re - distributed allocL h,w and allocH h,w to the remaining parts histRemL h,w and histRemH h,w to obtain new sub - histograms histL h,w ' and histH h,w ':

[0151] histL h，w ′(m) = histRemL h，w (m)+allocL h，w (m), 0 ≤ m < avg h，w

[0152] histH h，w ′(n) = histRemH h，w (n)+allocH h，w (n), 0 ≤ n < numBins - avg h，w (19)

[0153] S2.6. Within each block, perform histogram equalization on the histograms histL h，w ′ and histH h，w ′ within their respective gray - level ranges, mainly including:

[0154] (1) Calculate the probability density function pdfL h，w of the gray - level of histL h，w , and the formula is as follows:

[0155]

[0156] (2) Calculate histH h，w The probability density function pdfH of the gray levels of ′ h，w , the formula is as follows:

[0157]

[0158] (3) Accumulate the probability density function of pdfL h，w within its gray level range [0avg h，w - 1], and calculate the mapping function cdfL h，w , the formula is as follows:

[0159]

[0160] (4) Accumulate the probability density function of pdfH h，w within its gray level range [avg h，w numBins - 1], and calculate the mapping function cdfH h，w , the formula is as follows:

[0161]

[0162] (5) Combine cdfL h，w and cdfH h，w to obtain the final mapping function cdf h，w ;

[0163]

[0164] S3. According to the pixel position and the mapping function cdf h，w in each block, perform bilinear interpolation on the image to obtain the enhanced image y′.

[0165] The bilinear interpolation operation is a well - known and basic operation method in the field of image contrast enhancement. For details, refer to the paper Zuiderveld, Karel. "Contrast limited adaptive histogram equalization." Graphics gems IV. Academic Press Professional, Inc., 1994. In the present invention, the bilinear interpolation is only used as an operation method and will not be elaborated here.

[0166] It should be particularly emphasized that the scope of action of the above step S2 is within each image block, and the function in step S2 is to calculate a new mapping curve within each image block by the method of histogram piecewise reassignment.

[0167] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. For those skilled in the art, various changes and modifications can be made to the embodiments of the present invention. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A dual histogram equalization method based on image block partitioning, characterized in that, The method includes the following steps: S1, image block division; S2, operations on each sub-block of the image block division; that is, within each sub-block: S2.1, statistical histogram, calculate the histogram clipping threshold and the average gray level of the block; S2.2, divide the histogram into two sub-histograms according to the average gray level; S2.3, clip the two histograms respectively according to the clipping threshold; S2.4, allocate the clipped parts in the two sub-histograms according to a preset allocation method; S2.5, and superimpose them on the remaining sub-histograms respectively; S2.6, perform histogram equalization on the two sub-histograms within their respective gray level ranges, and combine the two obtained mapping curves as the final mapping function; S3, perform bilinear interpolation operation on the image according to the pixel position and the mapping function of each block to obtain the enhanced image.

2. A double histogram equalization method based on image block segmentation according to claim 1, characterized in that, The step S1 is to divide the input luminance image into blocks, and further includes: (1) Input the luminance image y, and the width and height of the input image are imgWidth and imgHeight respectively; (2) Set the number of vertical blocks bloVer and the number of horizontal blocks bloHor, where bloVer = 8 and bloHor = 8; if the block division method is vertical block division, then: sideVer = ((imgHeight / 2) / bloVer) × 2 (1) remVer = (imgHeight / 2) % bloVer (2) where sideVer represents the minimum number of pixels in the vertical block division, and remVer determines whether the pixel height of the image can be divisible by the number of vertical blocks during vertical block division; when remVer is not equal to 0, that is, in the process of block division, there is a non-divisible situation, the pixel height of the first remVer blocks in the vertical block division will be operated with +2, as shown in the following formula (3), to ensure that this allocation method can satisfy the following formula (4); that is (sideVer + 2) × remVer + sideVer × (bloVer - remVer) = imgHeight (4) In formula (3), bloHeight(h) represents the pixel height of the h-th block during vertical block division, h ∈ [1, bloVer]; similarly, in the process of horizontal block division: sideHor = ((imgWidth / 2) / bloHor) × 2 (5) remHor = (imgWidth / 2) % bloHor (6) (sideHor + 2) × remHor + sideHor × (bloHor - remHor) = imgWidth (8) In formulas (5)-(7), sideHor represents the minimum number of pixels in the horizontal block division, remHor determines whether the pixel width of the image can be divisible by the number of horizontal blocks during horizontal block division, bloWidth(w) represents the pixel width of the w-th block during horizontal block division, w ∈ [1, bloHor]; (3) Divide the input image y into bloVer × bloHor blocks according to the calculated results.

3. A dual histogram equalization method based on image block segmentation according to claim 1, characterized in that, The step S2 further includes: S2.1, within each block y h,w inside: (1) Statistically analyze the current block y h,w to obtain the histogram information hist h，w , with a total of numBins gray levels; (2) Calculate y h，w to obtain the average gray value avg h，w , and the formula is as follows: Among them represents rounding down, hist h，w (i) represents the number of pixels in y h，w with pixel value i, where i ∈ [0, mumBins); (3) Calculate the clipping threshold clipThre of this block according to the preset clipping ratio clipRatio h，w , and the calculation process is as follows: Here, clipRatio = 3; S2.2, within each block y h,w divide the histogram into two sub-histograms of high and low according to the average gray value avg h,w ; (1) Sub - histogram 1 is derived from the lower gray - level region of hist h，w and is denoted as histL h,w ; Sub - histogram 2 is derived from the higher gray - level region of hist h，w and is denoted as histH h,w , and the specific division method is as follows: histL h，w (m) = hist h，w (i), 0 ≤ m, i < avg h,w (11) histH h，w (n) = hist h，w (i), avg h，w ≤ i < numBins, 0 ≤ n < numBins - avg h，w (12) where m and n respectively represent the m-th and n-th gray levels of the sub-histograms histL h，w , histH h，w , m ∈ [0, avg h，w ), n ∈ [0, numBins - avg h，w ); (2) The two sub-histograms need to satisfy: S2.

3. Within each block, use clipThre obtained in (3) of step S2.1 h，w to clip histL h，w and histH h，w respectively. (1) Denote the remaining part as histRemL h，w and histRemH h，w , and the clipping process is as follows: (2) Count the number of pixels numL and numH cut out at each gray level, and accumulate them respectively to obtain the total number of pixels in the cut-out part, denoted as totalL h，w and totalH h，w : (3) This clipping process needs to satisfy: S2.4, within each block, re - allocate the cropped parts of the two sub - histograms according to a preset allocation method, and denote the two obtained sub - histograms as allocL h，w and allocH h，w : S2.5, within each block, add the reallocated allocL h，w and allocH h，w to the remaining parts histRemL h，w and histRemH h，w to obtain new sub-histograms histL h，w ' and histH h，w ': histL h，w ′(m) = histRemL h，w (m) + allocL h，w (m), 0 ≤ m < avg h,w histH h，w ′(n) = histRemH h，w (n) + allocH h，w (n), 0 ≤ n < numBins - avg h，w (19) S2.6, within each block, perform histogram equalization on the histograms histL h,w ′ and histH h,w ′ within their respective grayscale ranges, including: (1) Calculate histL h,w The probability density function pdfL of the gray level of h,w , the formula is as follows: (2) Calculate histH h,w The probability density function pdfH of the gray level of h,w , the formula is as follows: (3) Cumulatively add the probability density function of pdfL h,w within its grayscale range [0avg h，w -1], and calculate the mapping function cdfL h,w as follows: (4) Cumulatively add the probability density function of pdfH h,w within its grayscale range [avg h，w numBins - 1], and calculate the mapping function cdfH h，w as follows: (5) Combine cdfL h,w and cdfH h,w to obtain the final mapping function cdf h，w ; 4. A method for double histogram equalization based on image block segmentation according to claim 1, characterized in that, The scope of action of step S2 is in each image block, and the function in step S2 is to calculate a mapping function within each image block by the method of histogram piecewise redistribution.

5. A method for double histogram equalization based on image block segmentation according to claim 1, characterized in that In step S2.1, numBins = 256.

6. A method for dual histogram equalization based on image block segmentation according to claim 1, characterized in that, In the step S3, according to the pixel position and the mapping function cdf in each block h，w , a bilinear interpolation operation is performed on the image to obtain an enhanced image y'.

Images