CN-122023138-A - Multi-scale self-adaptive short wave infrared image enhancement method and device

CN122023138ACN 122023138 ACN122023138 ACN 122023138ACN-122023138-A

Abstract

The application provides a multi-scale self-adaptive short wave infrared image enhancement method and device, the method comprises the steps of calculating a cumulative probability density function of an original image histogram, determining a segmentation gray level threshold value for segmenting the original image histogram based on the cumulative probability density function and a golden section value, segmenting the original image histogram according to the segmentation gray level threshold value to obtain a first sub-image histogram and a second sub-image histogram, respectively carrying out multi-scale self-adaptive correction on the first sub-image histogram and the second sub-image histogram to obtain a first corrected histogram corresponding to the first sub-image histogram and a second corrected histogram corresponding to the second sub-image histogram, determining an enhancement histogram of the original image histogram according to the first corrected histogram and the second corrected histogram, and determining an enhancement image of the original image according to the enhancement histogram. To enhance the effect of image enhancement when image enhancement is performed.

Inventors

CHEN YUSHENG

Assignees

北京遥感设备研究所

Dates

Publication Date: 20260512
Application Date: 20251230

Claims (10)

1. The multi-scale self-adaptive short-wave infrared image enhancement method is characterized by comprising the following steps of: Calculating a cumulative probability density function of an original image histogram, wherein the original image histogram is a histogram of an original image, and the original image is a short wave infrared image; Determining a segmentation gray threshold for segmenting the original image histogram based on the cumulative probability density function and golden section values; Dividing the original image histogram according to the dividing gray threshold value to obtain a first sub-image histogram and a second sub-image histogram; respectively carrying out multi-scale self-adaptive correction on the first sub-image histogram and the second sub-image histogram to obtain a first corrected histogram corresponding to the first sub-image histogram and a second corrected histogram corresponding to the second sub-image histogram; Determining an enhancement histogram of the original image histogram from the first corrected histogram and the second corrected histogram; and determining an enhanced image of the original image according to the enhanced histogram.
2. The method of claim 1 wherein said determining a segmentation gray threshold for segmenting the original image histogram based on the cumulative probability density function and golden section values comprises: And taking the value closest to the golden section value in the cumulative probability density function as a segmentation gray level threshold value.
3. The method of claim 1, wherein the dividing the original image histogram according to the dividing gray threshold to obtain a first sub-image histogram and a second sub-image histogram comprises: taking a histogram formed by parts of the original image histogram, the gray level of which is not more than the segmentation gray level threshold value, as a first sub-image histogram; and taking a histogram formed by parts with gray levels larger than the segmentation gray threshold value in the original image histogram as a second sub-image histogram.
4. The method of claim 3, wherein performing multi-scale adaptive correction on the first sub-image histogram and the second sub-image histogram to obtain a first corrected histogram corresponding to the first sub-image histogram and a second corrected histogram corresponding to the second sub-image histogram, respectively, includes: determining an exposure of the original image histogram; Determining power law transformation parameters based on the exposure; aiming at the first sub-image histogram, acquiring a scale transformation coefficient corresponding to each scale adjustment strategy in a plurality of scale adjustment strategies to obtain a plurality of first scale transformation coefficients; Aiming at the second sub-image histogram, acquiring a scale transformation coefficient corresponding to each scale adjustment strategy in a plurality of scale adjustment strategies to obtain a plurality of second scale transformation coefficients; Correcting the first sub-image histogram based on the plurality of first scale conversion coefficients, the power law conversion parameters and a first minimum frequency and a first maximum frequency in the frequencies of all gray levels in the first sub-image histogram to obtain a first corrected histogram; and correcting the second sub-image histogram based on the second scale transformation coefficients, the power law transformation parameters and the second minimum frequency and the second maximum frequency in the frequencies of all gray levels in the second sub-image histogram to obtain a second corrected histogram.
5. The method of claim 4, wherein determining an enhanced histogram of the original image histogram from the first corrected histogram and the second corrected histogram comprises: Determining a first probability density function and a first cumulative probability density function of the first corrected histogram and a second probability density function and a second cumulative probability density function of the second corrected histogram; determining a first equalized transform function of the first corrected histogram based on the gray level threshold, the first probability density function, and a first cumulative probability density function, and determining a second equalized transform function of the second corrected histogram based on the gray level threshold, the second probability density function, and a second cumulative probability density function; performing equalization mapping on the first corrected histogram based on the first equalization transformation function to obtain a first mapping result, and performing equalization mapping on the second corrected histogram based on the second equalization transformation function to obtain a second mapping result; And combining the first mapping result and the second mapping result to obtain an enhanced histogram of the original image histogram.
6. The method of claim 4, wherein said determining the exposure of the original image histogram comprises: acquiring a plurality of gray levels of the original image histogram; determining, for each gray level of the plurality of gray levels, a probability density function corresponding to the gray level for the original image histogram; Calculating the product of the gray level and the probability density function to obtain the product corresponding to the gray level; Summing a plurality of products corresponding to the gray levels to obtain a summation result; the ratio of the result of the summation to the number of gray levels included in the plurality of gray levels is taken as an exposure.
7. The method of claim 4, wherein determining an enhanced image of the original image from the enhanced histogram comprises: and taking the image corresponding to the enhanced histogram as an enhanced image of the original image.
8. A multi-scale adaptive short wave infrared image enhancement device, comprising: The computing unit is used for computing a cumulative probability density function of an original image histogram, wherein the original image histogram is a histogram of an original image, and the original image is a short wave infrared image; A determining unit configured to determine a segmentation gray threshold for segmenting the original image histogram based on the cumulative probability density function and golden section values; the segmentation unit is used for segmenting the original image histogram according to the segmentation gray threshold value to obtain a first sub-image histogram and a second sub-image histogram; The correction unit is used for respectively carrying out multi-scale self-adaptive correction on the first sub-image histogram and the second sub-image histogram to obtain a first corrected histogram corresponding to the first sub-image histogram and a second corrected histogram corresponding to the second sub-image histogram; The determining unit is further configured to determine an enhancement histogram of the original image histogram according to the first corrected histogram and the second corrected histogram; The determining unit is used for determining an enhanced image of the original image according to the enhanced histogram.
9. An electronic device comprising a memory, a processor and a computer program stored in the memory and running on the processor, characterized in that the processor implements the steps of the method according to any of claims 1 to 7 when the computer program is executed.
10. A computer readable storage medium storing a computer program, characterized in that the computer program when executed by a processor implements the steps of the method according to any one of claims 1 to 7.

Description

Multi-scale self-adaptive short wave infrared image enhancement method and device Technical Field The application belongs to the technical field of image enhancement, and particularly relates to a multi-scale self-adaptive short-wave infrared image enhancement method and device. Background The short wave infrared camera has important application value in the fields of remote sensing monitoring, astronomical observation, environmental monitoring and the like, and the imaging quality of the short wave infrared camera directly influences the accuracy of target identification and information extraction. However, the existing short-wave infrared image (short-wave image) enhancement technology has the remarkable defects that the image is often subjected to the influence of a short-wave spectrum special effect and an imaging environment, the problems of low contrast, blurred details, obvious noise interference and the like are caused, and particularly, under a complex illumination or weak light scene, a large gap exists between the image and the actual application requirement. Although the traditional image enhancement method (such as histogram equalization HE) can improve contrast to a certain extent, local area enhancement and spectral information distortion are easy to occur because special effects (narrow wave bands and concentrated energy) of the short-wave image spectral distribution are not considered, and the contradiction between noise and detail preservation in different scenes cannot be adaptively processed. Although the improvement method is optimized for partial defects, single scale change cannot adapt to the differentiation requirement of the multi-frequency band information of the short-wave image, so that the enhancement effect of high-frequency details (such as edge textures) or low-frequency backgrounds (such as scene outlines) is insufficient. The above problems make it difficult to achieve high contrast and high fidelity image enhancement in complex scene applications of short wave cameras in the prior art, and a multi-scale adaptive enhancement scheme combining short wave spectrum characteristics is needed to solve the technical problem that the enhancement effect is poor when image enhancement is performed in the related art. Disclosure of Invention The application aims to provide a multi-scale self-adaptive short-wave infrared image enhancement method and device, which are used for solving the technical problem of poor image enhancement effect when image enhancement is carried out. In a first aspect of the embodiment of the present application, a multi-scale adaptive short-wave infrared image enhancement method is provided, including: Calculating a cumulative probability density function of an original image histogram, wherein the original image histogram is a histogram of an original image, and the original image is a short wave infrared image; Determining a segmentation gray threshold for segmenting the original image histogram based on the cumulative probability density function and golden section values; Dividing the original image histogram according to the dividing gray threshold value to obtain a first sub-image histogram and a second sub-image histogram; respectively carrying out multi-scale self-adaptive correction on the first sub-image histogram and the second sub-image histogram to obtain a first corrected histogram corresponding to the first sub-image histogram and a second corrected histogram corresponding to the second sub-image histogram; Determining an enhancement histogram of the original image histogram from the first corrected histogram and the second corrected histogram; and determining an enhanced image of the original image according to the enhanced histogram. In a second aspect of the embodiment of the present application, there is provided a multi-scale adaptive short-wave infrared image enhancement device, including: The computing unit is used for computing a cumulative probability density function of an original image histogram, wherein the original image histogram is a histogram of an original image, and the original image is a short wave infrared image; A determining unit configured to determine a segmentation gray threshold for segmenting the original image histogram based on the cumulative probability density function and golden section values; the segmentation unit is used for segmenting the original image histogram according to the segmentation gray threshold value to obtain a first sub-image histogram and a second sub-image histogram; The correction unit is used for respectively carrying out multi-scale self-adaptive correction on the first sub-image histogram and the second sub-image histogram to obtain a first corrected histogram corresponding to the first sub-image histogram and a second corrected histogram corresponding to the second sub-image histogram; The determining unit is further configured to determine an enhancement histogram of the original ima