A rapid quantization method for infrared measurement images of unpowered, weak flying targets
By using statistical grayscale histograms and equalization processing, the problems of information loss and unsatisfactory edge enhancement in the quantization of infrared images of unpowered, weak flying targets were solved, achieving complete preservation of image information and rapid quantization.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- CHINESE PEOPLES LIBERATION ARMY UNIT 92941
- Filing Date
- 2022-10-09
- Publication Date
- 2026-07-31
AI Technical Summary
Existing infrared image quantization methods cannot effectively preserve the effective information of unpowered and weak flying targets, and the edge enhancement effect is not ideal, which easily produces false edges and leads to a decrease in image quality.
By statistically analyzing the image's gray-level histogram, the original gray-level of pixels is equalized, and the gray-level range is projected into 256 levels. The quantization value of each gray level is calculated, and noise is processed by combining weights and thresholds to achieve rapid quantization.
It effectively preserves image information of unpowered and weak flying targets, enhances image quality, and enables rapid quantization processing.
Smart Images
Figure CN115797197B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of image measurement technology, specifically to a rapid quantization method for infrared measurement images of unpowered, weak, and flying targets. Background Technology
[0002] Current quantization methods for infrared images are all based on the statistical characteristics of the entire image. However, due to the low infrared radiation energy of small, unpowered aerial targets and the generally wide radiation range of infrared sensors, the effective signal range of the target in the infrared image is relatively narrow. If existing techniques are used to perform grayscale statistics on the entire image and scale its grayscale range, it will inevitably lead to the loss of effective information and cause image distortion. Another approach is to enhance the image edges. However, due to the radiation characteristics of infrared images, the gradient transformation at the edges of small target images is relatively gentle, resulting in unsatisfactory edge enhancement effects and a tendency to produce false image edges, which in turn eliminates the image texture of small targets and reduces image quality. Summary of the Invention
[0003] In view of this, the present invention provides a rapid quantization method for infrared measurement images of unpowered, weak flying targets, which can completely retain the effective information of the images of unpowered, weak flying targets and has a significant enhancement effect.
[0004] The rapid quantization method for infrared measurement images of unpowered, weak flying targets disclosed herein includes the following steps:
[0005] Step S1: Calculate the image grayscale histogram containing pixel count and location information;
[0006] Step S2, pixel original grayscale equalization;
[0007] Step S3: Project the grayscale range of the entire image into 256 grayscale levels and calculate the quantization value of each grayscale level.
[0008] Step S4: Based on the projection and quantization results from step S3, perform rapid quantization processing on the image.
[0009] Furthermore, the specific method in step S1 includes the following steps:
[0010] When calculating the grayscale values of the entire image, pixels with the same grayscale value and the total number of pixels will be counted. Greater than pixel grayscale values and weights Multiplication, as shown in the following formula:
[0011] if: ,and
[0012]
[0013] in, Indicates pixel position, ), ], Rows and columns for pixel regions with the same gray level; The pixel grayscale value; The weights represent the number of effective gray levels in the image, starting from 1 and increasing sequentially. The weight is The total number of pixels;
[0014] This is an empirical threshold set in advance based on the characteristics of the image. If the number of pixels at the same gray level is less than this value, it is considered noise by default.
[0015] Further, step S2 includes the following steps:
[0016] The original grayscale of the pixels is equalized, expanding the grayscale of pixels with fewer pixels and compressing the grayscale of pixels with more pixels: the total number of pixels in the image is... grayscale range is Then its first k The probability of each gray level appearing is ;
[0017] in, This represents the k-th gray level. In the image The number of pixels;
[0018] Image gray-level cumulative distribution function S k The function expression for histogram equalization of an image is:
[0019]
[0020] The corresponding inverse transform of this function, i.e., the function after histogram equalization, is:
[0021] .
[0022] Further, step S3 includes the following steps:
[0023] To improve quantization speed, the starting and ending grayscale values are statistically analyzed, the entire grayscale range is transformed into 256 grayscale levels, and the quantization value of each grayscale level is calculated, as follows:
[0024]
[0025] =
[0026] Where d is the dimension of quantification. These are the minimum and maximum gray values of the entire image after equalization, respectively. , This represents the quantization value corresponding to each gray level.
[0027] Further, step S4 includes the following steps:
[0028] To improve quantization speed, for values smaller than... The grayscale value is directly assigned zero, and the grayscale value is directly assigned 255, as shown in the following formula:
[0029] if ,but
[0030] if ,but ;
[0031] Other pixels undergo lookup table quantization, as shown in the following formula:
[0032] if
[0033] but
[0034] in, The grayscale of the image after equalization. Pixel positions in the entire image To quantize the grayscale index, To quantize grayscale, the correspondence between the original grayscale value and the 8-bit grayscale value quantized pixels is represented, thereby quantizing all pixels in the equalized image.
[0035] The infrared image quantization method for weak targets disclosed herein employs a method of stretching the grayscale of weak targets, compressing the grayscale levels of large pixel quantities, and increasing the grayscale levels of weak targets, which effectively enhances the grayscale levels of unpowered flying targets. At the same time, in order to improve quantization efficiency, it proposes to fix the ineffective grayscale levels and quantize the effective grayscale levels at 256 levels to obtain the quantization dimensions, thereby achieving rapid image quantization.
[0036] Compared with existing technologies, the beneficial effects of this disclosure are: ① It completely preserves and enhances the effective information of images of weak, unpowered flying targets; ② It enables rapid quantization; ③ The algorithm is simple and practical. Attached Figure Description
[0037] Figure 1 This is a flowchart of the present disclosure;
[0038] Figure 2 This is an example result after denoising, equalizing, and quantizing the original histogram. Detailed Implementation
[0039] The present invention will now be described in detail with reference to the accompanying drawings and embodiments.
[0040] This disclosure provides a method for quantizing infrared images of weak targets. (See attached image) Figure 1 As shown, an exemplary embodiment of this disclosure includes the following steps:
[0041] S1, statistically analyzes the image grayscale histogram containing pixel count information.
[0042] Since noise is mostly composed of isolated points, pixel location information is incorporated into histogram statistics to effectively remove it. When calculating the grayscale values and pixel location information of the entire image, pixels with the same grayscale value and a total number greater than a certain threshold are selected. pixel grayscale values and weights Multiplication, as shown in the following formula:
[0043] if: ,and
[0044]
[0045] in, Indicates pixel position, ), ], Rows and columns for pixel regions with the same gray level; The pixel grayscale value; The weights represent the number of effective gray levels in the image, starting from 1 and increasing sequentially. The weight is The total number of pixels;
[0046] To improve noise reduction, the weights of the first few pixels are set to negative integer values, and the statistic is ensured to be greater than zero when calculating the histogram. hour, ;
[0047] This is an empirical threshold set in advance based on the characteristics of the image. If the number of pixels at the same gray level is less than this value, it is considered noise by default.
[0048] S2, Pixel Original Grayscale Equalization
[0049] The original grayscale values of the pixels are equalized, expanding the grayscale values of pixels with fewer pixels and compressing the grayscale values of pixels with more pixels. The total number of pixels in the image is... grayscale range is Then its first The probability of each gray level appearing is , ;
[0050] in, Indicates the first k One gray level, In the image The number of pixels;
[0051] Image gray-level cumulative distribution function S k The function expression for histogram equalization of an image is:
[0052]
[0053] The corresponding inverse transform of this function, i.e., the function after histogram equalization, is:
[0054] .
[0055] S3, calculates the 8-bit gray quantization dimension.
[0056] To improve quantization speed, the starting and ending grayscale values are statistically analyzed, the entire grayscale range is transformed into 256 grayscale levels, and the quantization value of each grayscale level is calculated. The formula is as follows:
[0057]
[0058] =
[0059] Where d is the dimension of quantification. These are the minimum and maximum gray values of the entire image after equalization, respectively. , This represents the quantization value corresponding to each gray level.
[0060] S4, Fast Image Quantization Processing.
[0061] To improve quantization speed, for values smaller than... The grayscale value is directly assigned zero, and the grayscale value is directly assigned 255, as shown in the following formula:
[0062] if ,but
[0063] if ,but
[0064] Other pixels undergo lookup table quantization, as shown in the following formula:
[0065] if
[0066] but
[0067] in, The grayscale of the image after equalization. Pixel positions in the entire image To quantize the grayscale index, To quantize grayscale, the correspondence between the original grayscale value and the 8-bit grayscale value quantized pixels is represented, thereby quantizing all pixels in the equalized image.
[0068] The results of denoising, equalizing, and quantizing the example original histogram are shown in the attached figure. Figure 2 As shown.
[0069] In summary, the above are merely preferred embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.
Claims
1. A method for fast quantization of infrared measurement images of unpowered airborne targets, characterized in that, Includes the following steps: Step S1: Calculate the image grayscale histogram containing pixel count and location information; Step S2, pixel original grayscale equalization; Step S3: Project the grayscale range of the entire image into 256 grayscale levels and calculate the quantization value of each grayscale level. Step S4: Based on the projection and quantization results from step S3, perform rapid quantization processing on the image; The specific method in step S1 includes the following steps: When the gray values of the whole image are counted, the pixel gray values with the same gray and the total number of pixels greater than are multiplied by the weight value as follows: If: and in, Indicates pixel position, ), ], Rows and columns for pixel regions with the same gray level; The pixel grayscale value; The weight represents the number of effective gray levels in the image, starting from 1 and increasing sequentially. The weight is The total number of pixels; This is an empirical threshold set in advance based on the characteristics of the image. If the number of pixels at the same gray level is less than this value, it is considered noise by default.
2. The quantization method as described in claim 1, wherein step S2 includes the following steps: The original grayscale of the pixels is equalized, expanding the grayscale of pixels with fewer pixels and compressing the grayscale of pixels with more pixels: the total number of pixels in the image is... grayscale range is Then its first k The probability of each gray level appearing is ; in, Indicates the first k One gray level, In the image The number of pixels; Cumulative distribution function of gray scale of image S k That is, the function expression of histogram equalization of the image is: The corresponding inverse transform of this function, i.e., the function after histogram equalization, is: 。 3. The quantization method of claim 2, wherein, Step S3 includes the following steps: To improve quantization speed, the starting and ending grayscale values are statistically analyzed, the entire grayscale range is transformed into 256 grayscale levels, and the quantization value of each grayscale level is calculated, as follows: = where d is the dimension of quantization, respectively the minimum and maximum gray value of the whole image after equalization, , is the quantization value corresponding to each gray level.
4. The quantization method of claim 3, wherein, Step S4 includes the following steps: To improve the quantization speed, for less than gray scale directly assigned to zero, otherwise directly assigned to 255, as follows: If , then If , then ; Other pixels undergo lookup table quantization, as shown in the following formula: If then in, The grayscale of the image after equalization. Pixel positions in the entire image For quantizing grayscale indexes, To quantize grayscale, the correspondence between the original grayscale value and the 8-bit grayscale value quantized pixels is represented, thereby quantizing all pixels in the equalized image.