CN-121999247-A - Interference level evaluation method, device, equipment and medium for environment image

CN121999247ACN 121999247 ACN121999247 ACN 121999247ACN-121999247-A

Abstract

The application provides an interference level evaluation method, device, equipment and medium for an environment image, which are used for the technical field of interference evaluation of the environment image and can solve the technical problem of low efficiency and accuracy of the existing interference evaluation. The method comprises the steps of determining a plurality of corresponding target image characteristic parameters according to the real-time image to be evaluated and the corresponding real-time interference category thereof, carrying out weighted summation on the target image characteristic parameters according to the corresponding preset weight values to obtain corresponding target evaluation values, and determining the corresponding target interference level of the real-time image to be evaluated according to the target evaluation values and a plurality of evaluation thresholds, so that the efficiency and the accuracy of interference evaluation are improved.

Inventors

GAO CHI
JIA XINYIN
ZHANG ZHAOHUI
JIANG XIN
Hou pan
LI SIYUAN
HU BINGLIANG

Assignees

中国科学院西安光学精密机械研究所

Dates

Publication Date: 20260508
Application Date: 20260130

Claims (10)

1. A method for evaluating interference level of an environmental image, the method comprising: Determining a plurality of corresponding target image characteristic parameters according to the real-time image to be evaluated and the corresponding real-time interference category thereof, and carrying out weighted summation on the target image characteristic parameters according to the corresponding preset weight values to obtain corresponding target evaluation values; determining a target interference level corresponding to the image to be evaluated in real time according to the target evaluation value and a plurality of evaluation thresholds; The target image characteristic parameters are obtained based on the following steps: Calculating a plurality of image characteristic parameters corresponding to each target to-be-evaluated image according to the interference category of each target to-be-evaluated image, and classifying the target to-be-evaluated images with the same interference category based on the corresponding preset interference level to obtain a plurality of target image groups, wherein the interference category comprises smoke interference, camouflage net interference and false target interference; Calculating an inter-group mean square value and an intra-group mean square value corresponding to each image characteristic parameter based on a plurality of image characteristic parameters corresponding to each target image group, and obtaining a plurality of target characteristic values corresponding to the interference category according to the division value between the inter-group mean square value and the intra-group mean square value; And comparing each target characteristic value with a corresponding preset threshold value, and determining the target image characteristic parameters corresponding to each interference category according to the comparison result.
2. The method according to claim 1, wherein calculating a plurality of image feature parameters corresponding to each other according to the interference category of each target image to be evaluated comprises: Aiming at the target image to be evaluated of the smoke interference, calculating a corresponding divisor according to the gray standard deviation corresponding to each of the standard image and the target image to be evaluated to obtain an image contrast attenuation coefficient; based on an image gray level map corresponding to the target image to be evaluated, extracting corresponding edge pixels according to a Canny edge detection algorithm, and calculating edge information entropy according to gray distribution entropy values corresponding to the edge pixels; performing pixel-by-pixel calculation based on the standard image to obtain a color reference image, and calculating color difference between the color reference image and the target image to be evaluated based on CIE Lab color space; Performing size division on the image gray level graph to obtain a plurality of target gray level squares and non-target gray level squares, and performing linear fitting on the target gray level squares and the non-target gray level squares to obtain an image fractal dimension; And determining a plurality of image characteristic parameters based on the image contrast attenuation coefficient, the edge information entropy, the chromatic aberration and the image fractal dimension.
3. The method according to claim 1, wherein calculating a plurality of image feature parameters corresponding to each other according to the interference category of each target image to be evaluated comprises: Aiming at the target to-be-evaluated image interfered by the camouflage net, calculating a gray level co-occurrence matrix according to the distribution relation of gray level space in the target to-be-evaluated image; calculating a shape feature comprehensive difference value corresponding to the target image to be evaluated based on Fourier transformation; carrying out gray scale division on an edge image corresponding to the target image to be evaluated to obtain a plurality of image gray levels, and calculating an energy spectrum according to the image gray levels to obtain an edge characteristic comprehensive difference value; according to the component difference values corresponding to the standard image and the target image to be evaluated in brightness, contrast and structure, carrying out weighted summation on the component difference values to obtain the structural similarity; Extracting the energy difference between the standard image and the target image to be evaluated in a high frequency band according to Fourier transformation, and calculating high-frequency energy detail characteristics; and determining a plurality of image characteristic parameters according to the gray level co-occurrence matrix, the shape characteristic comprehensive difference value, the edge characteristic comprehensive difference value, the structural similarity and the high-frequency energy detail characteristic.
4. The method according to claim 1, wherein calculating a plurality of image feature parameters corresponding to each other according to the interference category of each target image to be evaluated comprises: Calculating pixel values of the target to-be-evaluated image under each HSV channel based on HSV space aiming at the target to-be-evaluated image interfered by the false target, and calculating corresponding color consistency difference values according to the pixel values; Based on gray level images corresponding to the standard image and the target image to be evaluated, extracting texture histograms corresponding to the standard image and the target image to be evaluated, and calculating information entropy differences to obtain texture entropy; Determining a corresponding edge direction histogram and an edge curvature histogram according to an edge image corresponding to the target image to be evaluated, and calculating an edge characteristic difference value according to a difference value between the edge direction histogram and the edge curvature histogram; calculating local contrast distribution entropy between the standard image and the target image to be evaluated according to a sliding window; and determining a plurality of image characteristic parameters according to the color consistency difference value, the texture entropy, the edge characteristic difference value and the local contrast distribution entropy.
5. The method according to claim 1, wherein determining the target image feature parameter corresponding to each interference category according to the comparison result includes: determining the corresponding comparison result; if the comparison result shows that the value corresponding to the target characteristic value is larger than the preset threshold value, determining the image characteristic parameter corresponding to the target characteristic value as the target image characteristic parameter; And if the comparison result is that the value corresponding to the target characteristic value is not greater than the preset threshold value, determining that the image characteristic parameter corresponding to the target characteristic value is not the target image characteristic parameter.
6. The method of claim 1, wherein determining the target interference level corresponding to the real-time image to be evaluated according to the target evaluation value and a plurality of evaluation thresholds comprises: Sorting the plurality of evaluation thresholds according to the corresponding numerical values to obtain an evaluation threshold sequence, and determining a plurality of corresponding evaluation intervals and interference levels corresponding to the evaluation intervals according to the evaluation threshold sequence; And determining the target interference level according to the evaluation interval and the interference level.
7. The method of claim 6, wherein said determining said target interference level based on said evaluation interval and said interference level comprises: And determining a target evaluation interval corresponding to the value of the target evaluation value according to the value range corresponding to each evaluation interval, and determining the target interference level according to the interference level corresponding to the target evaluation interval.
8. An interference level assessment device for an environmental image, the device comprising: The weighting module is used for determining a plurality of corresponding target image characteristic parameters according to the real-time image to be evaluated and the corresponding real-time interference category thereof, and carrying out weighted summation on the target image characteristic parameters according to the corresponding preset weight values to obtain corresponding target evaluation values; The determining module is used for determining a target interference level corresponding to the image to be evaluated in real time according to the target evaluation value and a plurality of evaluation thresholds; the target image characteristic parameters are obtained based on the following modules: The system comprises a classification module, a classification module and a classification module, wherein the classification module is used for calculating a plurality of image characteristic parameters corresponding to each target to-be-evaluated image according to the interference category of each target to-be-evaluated image, classifying the target to-be-evaluated images with the same interference category based on the corresponding preset interference level to obtain a plurality of target image groups, wherein the interference category comprises smoke interference, camouflage network interference and false target interference; The computing module is used for computing an inter-group mean square value and an intra-group mean square value corresponding to each image characteristic parameter based on a plurality of image characteristic parameters corresponding to each target image group, and obtaining a plurality of target characteristic values corresponding to the interference category according to the division value between the inter-group mean square value and the intra-group mean square value; and the comparison module is used for comparing each target characteristic value with a corresponding preset threshold value, and determining the target image characteristic parameters corresponding to each interference category according to the comparison result.
9. A computer device, comprising: One or more processors; A memory; One or more programs, wherein the one or more programs are stored in memory and configured to be executed by the one or more processors, the one or more programs configured to perform the method of any of claims 1-7.
10. A computer readable storage medium having stored therein program code which is callable by a processor to perform the method according to any one of claims 1 to 7.

Description

Interference level evaluation method, device, equipment and medium for environment image Technical Field The present application relates to the field of interference evaluation technologies of environmental images, and in particular, to an interference level evaluation method, apparatus, device, and medium for an environmental image. Background The photoelectric measurement relies on visible light image characteristics (such as brightness, texture, edges and structures) of the target to perform parameter calculation, and interference such as smoke, camouflage net, false target and the like directly distorts the characteristics, such as smoke reduction contrast, camouflage net blurring of the target and background boundary, so that if corresponding interference levels are not identified, measurement results deviate from true values, such as misjudgment of the size, position or performance index of the target, and interference level assessment can provide basis for correction of the measurement results. The existing interference level evaluation method generally uses a transmissivity meter, a visibility meter and other devices to measure the attenuation coefficient or the transmissivity of aerosol such as smoke to light with specific wavelength, however, the method only reflects energy attenuation and completely ignores distortion of interference to image structures, textures and colors. Therefore, the existing interference level evaluation method has the problem of low evaluation efficiency and accuracy. Disclosure of Invention The application provides an interference level evaluation method, device, equipment and medium for environmental images, which are used for solving the problem that the existing interference level evaluation method has low evaluation efficiency and accuracy. In a first aspect, the present application provides a method for evaluating interference level of an environmental image, the method comprising: Determining a plurality of corresponding target image characteristic parameters according to the real-time image to be evaluated and the corresponding real-time interference category thereof, and carrying out weighted summation on the target image characteristic parameters according to the corresponding preset weight values to obtain corresponding target evaluation values; Determining a target interference level corresponding to the image to be evaluated in real time according to the target evaluation value and a plurality of evaluation thresholds; The target image characteristic parameters are obtained based on the following steps: calculating a plurality of image characteristic parameters corresponding to each target to-be-evaluated image according to the interference category of each target to-be-evaluated image, and classifying the target to-be-evaluated images with the same interference category based on the preset interference level corresponding to each target to obtain a plurality of target image groups; Calculating an inter-group mean square value and an intra-group mean square value corresponding to each image characteristic parameter based on a plurality of image characteristic parameters corresponding to each target image group, and obtaining a plurality of target characteristic values corresponding to interference categories according to the divisors between the inter-group mean square values and the intra-group mean square values; and comparing each target characteristic value with a corresponding preset threshold value, and determining target image characteristic parameters corresponding to each interference category according to the comparison result. In some embodiments of the present application, calculating a plurality of image feature parameters according to interference categories of respective target images to be evaluated includes: aiming at a target to-be-evaluated image of smoke interference, calculating a corresponding divisor according to gray standard deviations corresponding to the standard image and the target to-be-evaluated image respectively to obtain an image contrast attenuation coefficient; Based on an image gray level map corresponding to the target image to be evaluated, extracting corresponding edge pixels according to a Canny edge detection algorithm, and calculating edge information entropy according to gray distribution entropy values corresponding to the edge pixels; Performing pixel-by-pixel calculation based on the standard image to obtain a color reference image, and calculating color difference between the color reference image and the target image to be evaluated based on CIE Lab color space; performing size division on the image gray level graph to obtain a plurality of target gray level squares and non-target gray level squares, and performing linear fitting on the target gray level squares and the non-target gray level squares to obtain an image fractal dimension; a plurality of image feature parameters are determined based on the image c