Improved histogram equalization infrared image enhancement method

Abstract

The invention provides an improved histogram equalization infrared image enhancement method, and the method specifically comprises the following steps: S1, obtaining an original image, and carrying out the partitioning of the original image, and obtaining a plurality of image blocks; s2, performing histogram statistics on the original image to obtain a Gaussian smooth histogram; s3, calculating the scene complexity of the image blocks; s4, calculating an upper platform threshold value and a lower platform threshold value of the image block; s5, processing the Gaussian smooth histogram to obtain a corrected histogram; s6, calculating a cumulative distribution function according to the corrected histogram, and mapping the cumulative distribution function to a preset gray scale range to obtain a gray scale mapping function; s7, the image blocks are substituted into a gray mapping function for mapping, and an output image is obtained; different upper and lower platforms are utilized to control noise suppression or contrast stretching of an image block, so that the effects of noise suppression of a uniform region of an image and contrast stretching of a rich scene are achieved, and parameter adjustment is not needed, and the method belongs to the technical field of image processing.

Description

Technical Field

[0001] This invention belongs to the field of image processing technology, specifically relating to an improved method for infrared image enhancement using histogram equalization. Background Technology

[0002] With the increasing maturity of infrared focal plane arrays, they have advantages such as good concealment, strong resistance to electronic interference, and all-weather operation, and are widely used in military and civilian fields, such as military target detection, target recognition, target tracking and security control, as well as civilian intelligent transportation and security monitoring. However, due to the limitations of the detector's own performance, the generated infrared images have problems such as overall darkness, low contrast, unclear details and poor visual effects. Therefore, there is an urgent need for an infrared image enhancement technology to improve the quality of infrared images.

[0003] Most existing image enhancement methods are based on histograms, such as platform histogram equalization. These methods limit excessive background enhancement, suppress noise, and improve the dynamic range of the image by restricting the maximum number of gray levels in an image. While these algorithms can improve the visualization of infrared images under certain conditions, direct use of them can easily lead to excessive noise in uniform regions and oversaturation in some areas. In addition, selecting the platform threshold is also a challenge. In summary, current image enhancement algorithms cannot effectively balance image contrast enhancement and noise suppression. This invention limits the advantages of histogram equalization and improves upon them, proposing a balanced image enhancement algorithm. Summary of the Invention

[0004] The purpose of this invention is to provide an improved infrared image enhancement method using histogram equalization. By utilizing different upper and lower platforms to control noise suppression or contrast stretching of image blocks, it achieves the effects of noise suppression in uniform areas of the image and contrast stretching in rich scenes. Moreover, the platform is adaptive and requires no parameter tuning.

[0005] To achieve the above objectives, the present invention provides an improved histogram equalization method for infrared image enhancement, comprising the following steps: S1. Obtain the original image and the original image Perform block processing to obtain multiple image blocks; S2, For the original image Perform histogram statistics, and then apply Gaussian smoothing to the histogram to obtain the Gaussian smoothed histogram. S3. Calculate the scene complexity of the image patch; S4. Calculate the upper platform threshold of the image patch. With the lower platform threshold ; S5. Perform equalization processing on the histogram obtained after Gaussian smoothing in step S2 to obtain the corrected histogram. S6. Calculate the cumulative distribution function based on the modified histogram, and map the cumulative distribution function to the preset display grayscale range to obtain the grayscale mapping function; S7. Input the image blocks into the grayscale mapping function for mapping, and then stitch the mapped image blocks together to obtain the final output image.

[0006] Furthermore, the original image in step S1 The number of blocks is 9.

[0007] Furthermore, in step S3, the scene complexity of the image patch satisfies the formula: (1), in formula (1), Coefficient of variation The standard deviation of the image patch is represented. The mean of the image patch; the standard deviation of the image patch. The calculation formula is as follows: In the formula: The total number of pixel blocks. This represents the pixel grayscale value.

[0008] Furthermore, in step S4, the upper platform threshold T satisfies the formula: (4), in formula (4), This is the initial value for the platform; In step S4, the lower platform threshold L satisfies the formula: (5), in formula (5), This is the initial value for the lower platform. .

[0009] Furthermore, step S5 specifically includes the following steps: S51, Passing the threshold Limit the number of pixels at each gray level in the histogram, ensuring that the number of pixels at each gray level is greater than a certain threshold. The pixels are set to Pixel settings Satisfying the formula: (6) in, This represents the plateau threshold under the histogram, which can adjust the uniformity of the image. This is a configurable parameter, and its default value is 5. For Gaussian smoothed histograms, ; S52. Allocate the number of pixels exceeding the threshold to those below the threshold. The gray levels are allocated according to the following formula: (7), In formula (7), The number of pixels at the gray level of the histogram is greater than the threshold. The total number of pixels is given by formula (8): , (8).

[0010] Furthermore, the cumulative distribution function in step S6 is: (9), in formula (9) This represents the cumulative distribution function value.

[0011] Furthermore, the preset display grayscale range in step S6 is specifically 16-235, and the grayscale mapping function is specifically: (10).

[0012] Furthermore, bilinear interpolation is used for splicing in step S7.

[0013] This invention provides an improved infrared image enhancement method with histogram equalization, which is not a simple combination of obvious features. It organically combines block processing, histogram smoothing, scene complexity analysis, adaptive threshold calculation, and contrast-limited equalization to produce a synergistic effect: 1+1>2. It also achieves the technical effect of adaptively balancing contrast and noise suppression. The "previous frame histogram" and "Gaussian smoothing" processing introduced to meet engineering practice (FPGA real-time processing) further enhances its creativity and practicality.

[0014] The present invention will now be described in detail with reference to the accompanying drawings and embodiments. Attached Figure Description

[0015] Figure 1 This is a schematic diagram of the improved histogram equalization infrared image enhancement method.

[0016] Figure 2a This is a diagram illustrating the image processed by the traditional CLAHE algorithm. Figure 1 .

[0017] Figure 2b This is a schematic diagram of an image processed by an improved histogram equalization infrared image enhancement method. Figure 1 .

[0018] Figure 3a This is a schematic diagram of an image processed by the traditional CLAHE algorithm.

[0019] Figure 3b This is a schematic diagram of an image processed by an improved histogram equalization infrared image enhancement method. Detailed Implementation

[0020] To further illustrate the technical means and effects of the present invention in achieving its intended purpose, the specific implementation methods, structural features and effects of the present invention are described in detail below with reference to the accompanying drawings and embodiments.

[0021] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0022] In the description of this invention, it should be understood that the terms "center", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "aligned", "overlapping", "bottom", "inner", "outer", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this invention.

[0023] The terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature; in the description of this invention, unless otherwise stated, "a plurality of" means two or more. Example

[0024] This embodiment provides, for example Figure 1 Figure 3 shows an improved histogram equalization infrared image enhancement method, which includes the following steps: S1. Obtain the original image and the original image Perform block processing to obtain multiple image blocks; S2, For the original image Perform histogram statistics, and then apply Gaussian smoothing to the histogram to obtain the Gaussian smoothed histogram. S3. Calculate the scene complexity of the image patch; S4. Calculate the upper platform threshold of the image patch. With the lower platform threshold ; S5. The Gaussian smoothed histogram obtained in step S2 is subjected to equalization processing to obtain a contrast-limited plateau histogram. In this way, the plateau on the histogram is increased, the contrast of the uniform surface is limited, and the noise suppression effect is better. S6. Calculate the cumulative distribution function based on the modified histogram, and map the cumulative distribution function to the preset display grayscale range to obtain the grayscale mapping function; S7. Input the image blocks into the grayscale mapping function for mapping, and then stitch the mapped image blocks together to obtain the final output image.

[0025] Furthermore, the image in step S1 It is divided into 9 sections. The reason for dividing it into 9 sections is that compared to 4 sections, 9 sections have a stronger contrast ratio. Compared to 16 sections, it consumes less hardware resources and has less contrast loss.

[0026] Furthermore, in step S2, to meet engineering practice requirements: FPGA real-time processing, the histogram of the previous frame image is used for the current frame image processing, and the histogram is Gaussian smoothed. The histogram of the previous frame is used for processing the current frame. To prevent image flicker, the histogram is Gaussian smoothed using the following formula (template radius a=1). ). Where, input x is the histogram of the original image after processing; Furthermore, in step S3, the scene complexity of the image patch satisfies the formula: (1), in formula (1), The standard deviation of an image patch is expressed by the following formula: In the formula: The total number of pixel blocks. For pixel grayscale values, This represents the average value of the pixel block.

[0027] Formula (1) is the coefficient of variation. The calculation formula, and through the coefficient of variation To characterize the complexity of image patches; To perform normalization, the coefficient of variation was... The value is restricted, that is , (Empirical value) (2), normalize it, and then invert it to get: (3), The larger the value, the richer the scenarios. The smaller the value, the more uniform the scene, and the value is between 0 and 1.

[0028] Furthermore, in step S4, the upper platform threshold T satisfies the formula: (4), in formula (4), , which is the initial value for the platform; the platform threshold. The smaller the value, the smaller the proportion of background grayscale, and the stronger the effect of suppressing background noise in the image. Formula (4) is the upper platform threshold. The adaptive formula; In step S4, the lower platform threshold L satisfies the formula: (5), in formula (5), This is the initial value for the lower platform. Lower platform threshold The larger the value, the more it will compress the proportion of background grayscale, thereby suppressing background noise. Formula (5) is the lower platform threshold. The adaptive formula; Step S4 provides the calculation formulas for the upper platform threshold T and the lower platform threshold L, and it establishes the coefficients. The mathematical relationship between the two platform thresholds enables "adaptive" automation, with initial parameters... and It is usually a fixed constant, which can be applied to various scenarios after being set, without the need for manual parameter adjustment, which is a great advantage.

[0029] Furthermore, step S5 specifically includes the following steps: S51, Passing the threshold Limit the number of pixels at each gray level in the histogram, ensuring that the number of pixels at each gray level exceeds a threshold. The pixels are set to Pixel settings Satisfying the formula: (6) Among them, This represents the plateau threshold under the histogram, which can adjust the uniformity of the image. This is a configurable parameter, and its default value is 5. For Gaussian smoothed histograms, ; Increase less than the threshold Greater than The number of pixels in the histogram improves the contrast of the target.

[0030] S52. Allocate the number of pixels exceeding the threshold to those below the threshold. The gray levels are allocated according to the following formula: (7), In formula (7), The number of pixels at the gray level of the histogram is greater than the threshold. The total number of pixels is given by formula (8): , (8).

[0031] Steps S51 and S52 describe how to use a threshold The paper describes how to "crop" the histogram and redistribute the cropped pixels evenly across all gray levels, defining a specific implementation of "contrast-limited platform histogram equalization." Cropping limits over-enhancement, while average redistribution ensures the continuity of gray levels and avoids information loss. Furthermore, the cumulative distribution function in step S6 is: (9), in formula (9), This represents the cumulative distribution function value.

[0032] Furthermore, the preset display grayscale range in step S6 is specifically 16~235 (the grayscale range of a typical monitor is 16~235), and the grayscale mapping function is specifically: (10); This is an optimization measure for video display standards. In the standard, 16 represents black level, 235 represents white level, and the ranges of 0~15 and 236~255 are reserved for synchronization signals, etc., so that the enhanced image can be directly used for standard video display devices, avoiding signal clipping or distortion, and improving the practicality and compatibility of the algorithm.

[0033] Furthermore, bilinear interpolation is used for stitching in step S7. This invention clarifies that the interpolation algorithm used during stitching is bilinear interpolation. Bilinear interpolation is used to eliminate the block effect that may be caused by block processing, making the transition between image blocks smooth and natural. At the same time, it achieves a good balance between computational complexity and image quality, making it very suitable for real-time processing systems. It effectively improves the visual effect of the final output image, making the output image look like a coherent whole, rather than a pieced-together fragment.

[0034] Figure 2a , Figure 3a These are all output images from the traditional CLAHE algorithm. Figure 2b , Figure 3b These are all schematic diagrams of images processed by the improved histogram equalization infrared image enhancement method provided in this embodiment. Figure 2a Images and Figure 2b Image comparison, Figure 3a Images and Figure 3b The image comparison clearly shows differences in contrast, overexposure, and sky noise. Figure 2b Images, Figure 3b The image quality is better.

[0035] The above description, in conjunction with specific preferred embodiments, provides a further detailed explanation of the present invention. It should not be construed that the specific implementation of the present invention is limited to these descriptions. For those skilled in the art, various simple deductions or substitutions can be made without departing from the concept of the present invention, and all such modifications and substitutions should be considered within the scope of protection of the present invention.

Claims

1. An improved infrared image enhancement method using histogram equalization, characterized in that: Includes the following steps: S1. Obtain the original image and the original image Perform block processing to obtain multiple image blocks; S2, For the original image Perform histogram statistics, and then apply Gaussian smoothing to the histogram to obtain the Gaussian smoothed histogram. S3. Calculate the scene complexity of the image patch; S4. Calculate the upper platform threshold of the image patch. With the lower platform threshold ; S5. Perform equalization processing on the histogram obtained after Gaussian smoothing in step S2 to obtain the corrected histogram. S6. Calculate the cumulative distribution function based on the modified histogram, and map the cumulative distribution function to the preset display grayscale range to obtain the grayscale mapping function; S7. Input the image blocks into the grayscale mapping function for mapping, and then stitch the mapped image blocks together to obtain the final output image.

2. The improved infrared image enhancement method using histogram equalization as described in claim 1, characterized in that: The original image in step S1 The number of blocks is 9.

3. The improved infrared image enhancement method using histogram equalization as described in claim 1, characterized in that: In step S3, the scene complexity of the image patch satisfies the formula: (1), in formula (1), Coefficient of variation The standard deviation of the image patch is represented. The mean of the image patch; the standard deviation of the image patch. The calculation formula is as follows: In the formula: The total number of pixel blocks. This represents the pixel grayscale value.

4. An improved infrared image enhancement method using histogram equalization as described in claim 3, characterized in that: In step S4, the upper platform threshold T satisfies the formula: (4), in formula (4), This is the initial value for the platform; In step S4, the lower platform threshold L satisfies the formula: (5), in formula (5), This is the initial value for the lower platform. .

5. An improved infrared image enhancement method using histogram equalization as described in claim 1, characterized in that: Step S5 specifically includes the following steps: S51, Passing the threshold Limit the number of pixels at each gray level in the histogram, ensuring that the number of pixels at each gray level is greater than a certain threshold. The pixels are set to Pixel settings Satisfying the formula: （6） in, This represents the plateau threshold under the histogram, which can adjust the uniformity of the image. This is a configurable parameter, and its default value is 5. For Gaussian smoothed histograms, ; S52. Allocate the number of pixels exceeding the threshold to those below the threshold. The gray levels are allocated in a manner that satisfies the allocation formula: （7）， In formula (7), The number of pixels at the gray level of the histogram is greater than the threshold. The total number of pixels is given by formula (8): ， （8）。 6. An improved infrared image enhancement method using histogram equalization as described in claim 1, characterized in that: The cumulative distribution function mentioned in step S6 is: (9), in formula (9) This represents the cumulative distribution function value.

7. An improved infrared image enhancement method using histogram equalization as described in claim 1, characterized in that: The preset display grayscale range in step S6 is specifically 16-235, and the grayscale mapping function is specifically: (10).

8. An improved infrared image enhancement method using histogram equalization as described in claim 1, characterized in that: In step S7, bilinear interpolation is used for splicing.

Images