Image sharpness evaluation method for long focal length visible light imaging system

Abstract

The present application belongs to the technical field of image processing, and particularly relates to a kind of image sharpness evaluation methods for long focal length visible light imaging system. The present application first adopts median filter algorithm to pre-process image, and the algorithm is a typical representative of nonlinear filtering algorithm, which can eliminate isolated noise points, and can effectively protect edge contour information while removing random noise. This feature is very consistent with the high demand of edge information for sharpness function. Then, the pre-processed image is decomposed by NSST (non-subsampled shearlet transform), to obtain low-frequency components and high-frequency subbands, and a robust image sharpness evaluation method is constructed by combining direction consistency and contrast normalization.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of image processing, and particularly relates to an image definition evaluation method for a long-focus visible light imaging system. BACKGROUND

[0002] Compared with satellite reconnaissance, aerial reconnaissance has the characteristics of flexibility, maneuverability, accuracy and strong pertinence, and becomes a basic means for obtaining tactical intelligence in modern war. In recent years, aerial photoelectric pod gradually develops towards small size, single load, large aperture, long focal length and multi-spectral load, and the optical system also develops from a transmission system to a Cassegrainian catadioptric system, and long-range sea reconnaissance is also a capability that a high-performance photoelectric pod must have. In order to realize long-distance detection and recognition and high target resolution, the focal length of a visible light camera must be designed to be long enough. However, the long-focus and large-aperture optical imaging system is very sensitive to environmental changes, and small changes in temperature and atmospheric pressure will cause the defocus of the optical imaging load. Therefore, the long-focus aerial optical load must have an automatic focusing function, and the automatic focusing effect in different working conditions and different scenes also affects the image effect of aerial reconnaissance. At present, the automatic focusing method used by the aerial photoelectric pod is the image method, and the working principle of automatic focusing is that the focusing motor drives the focusing group, the potentiometer or the encoder to move at the same time, and the video processing unit at the rear end of the detector calculates the image definition value of the focusing group at different positions through the image definition evaluation function. After the focusing motor completes the full stroke traversal, it is located at the clearest position determined by the image definition function, thereby completing the automatic focusing work flow. The image method depends on the image definition evaluation function, and typical definition evaluation methods mainly include spatial domain type, frequency domain type, information entropy type, statistical type and model learning type. In actual engineering application, the most commonly used method is the spatial domain method, which is based on gradient calculation and has the advantages of low calculation complexity and being suitable for embedded real-time systems, but it is severely dependent on the contrast of the image. The action distance of the long-focus visible light imaging system covers several kilometers to hundreds of kilometers, and is subject to the diffraction limit and aperture limit of the optical system, the scattering and absorption of the atmosphere, the external working environment and other factors. With the increase of reconnaissance distance, the target contrast decreases, which puts forward higher requirements for the image definition evaluation method, and a definition evaluation function with strong robustness is needed to enable the long-focus optical system to realize accurate automatic focusing at different action distances.

[0003] The existing definition evaluation methods, such as gradient function and Laplacian function, have low sensitivity in long-distance weak texture or low contrast scenes, and may cause an increase in the failure probability of automatic focusing. SUMMARY

[0004] Therefore, the present application aims to provide a long-focus visible light imaging system-oriented image sharpness evaluation method to solve the problem that the sensitivity of the existing sharpness evaluation method decreases in a long-distance weak-texture or low-contrast scene, and the failure probability of automatic focusing increases. The present application can consider multi-scale features and direction information, has a weight adjustment mechanism, can improve the evaluation accuracy and robustness of the long-focus visible light imaging system in a complex environment, and effectively improves the accuracy of automatic focusing.

[0005] To achieve the above-mentioned purpose, the technical scheme of the present application is as follows:

[0006] A long-focus visible light imaging system-oriented image sharpness evaluation method, specifically comprising the following steps:

[0007] S1: The focusing motor of the long-focus visible light imaging system drives the focusing lens to move to obtain a focusing image;

[0008] S2: The focusing image is preprocessed to obtain a preprocessed image, and the preprocessed image is sequentially subjected to global linear normalization and local contrast normalization to obtain a normalized image;

[0009] S3: The normalized image is subjected to NSST decomposition to obtain high-frequency sub-band coefficients of different scales and different directions;

[0010] S4: Based on the high-frequency sub-band coefficients of different scales and different directions, the direction consistency of different scales is calculated;

[0011] S5: Based on the direction consistency of different scales, a local sharpness function of different scales and different directions is constructed;

[0012] S6: The local sharpness values of different scales and different directions are weighted and fused to obtain a fused image;

[0013] S7: The pixel values corresponding to each pixel point of the fused image are averaged to obtain the sharpness evaluation value of the focusing image.

[0014] Further, in step S2, the preprocessing operation is specifically a median filtering operation.

[0015] Further, in step S2, the calculation formula used for the global linear normalization and the local contrast normalization of the preprocessed image is:

[0016] ;

[0017] ;

[0018] wherein, is the preprocessed image, For globally linearly normalized images, The minimum pixel value in the preprocessed image. The maximum pixel value in the preprocessed image. To prevent zero items, The coordinates of the pixel. For locally contrast-normalized images, To obtain the average value of pixels within a local window W, This is to obtain the standard deviation of pixels within a local window W.

[0019] Furthermore, in step S4, the calculation formula used to calculate the directional consistency at different scales is as follows:

[0020] ;

[0021] in, For directional consistency at the s-scale, For the first Scale, First High-frequency subband coefficients in each direction.

[0022] Furthermore, the expressions for the local sharpness function at different scales and in different directions are as follows:

[0023] ;

[0024] in, For the first Scale, First Local sharpness values ​​in each direction, For directional consistency at the s-th scale, For the first Scale, First High-frequency subband coefficients in each direction.

[0025] Furthermore, the calculation formula used for weighted fusion of local sharpness values ​​at different scales and in different directions is as follows:

[0026] ;

[0027] ;

[0028] ;

[0029] in, For the first Scale, First Local sharpness values ​​in each direction, For the first Scale, First Weight coefficients for each direction, For the first scale, the subband energy of the fused image.

[0030] Further, the calculation formula for taking the average of the pixel values corresponding to each pixel point of the fused image is:

[0031]

[0032] Wherein, Score is the definition evaluation value.

[0033] Compared with the prior art, the present application can achieve the following beneficial effects:

[0034] The image definition evaluation method for the long focal length visible light imaging system of the present application adopts median filtering as a preprocessing means, removes noise interference while retaining image detail information to the greatest extent, performs contrast normalization processing on the image, improves the adaptability of the evaluation method to the environment, introduces direction consistency measurement at the subband level through NSST processing on the image, combines it with the subband energy according to an adaptive weight, can accurately distinguish the real blur from the energy changes caused by exposure and contrast difference, is more sensitive to the images (weak texture, low contrast) obtained by the long focal length optical system at a long distance, and can effectively assist the focusing system to complete accurate automatic focusing. BRIEF DESCRIPTION OF DRAWINGS

[0035] The accompanying drawings, which form a part of this application, are included to provide a further understanding of the application and are incorporated in and constitute a part of this application. The drawings illustrate embodiments of the present application and, together with the description, serve to explain the present application. In the drawings:

[0036] Figure 1 The flowchart of the image definition evaluation method for the long focal length visible light imaging system according to the present application. DETAILED DESCRIPTION

[0037] In order to make the purpose, technical scheme and advantages of the present application clearer, the present application will be further described in detail below with reference to the drawings and specific embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application, and do not constitute a limitation on the present application.

[0038] It should be noted that the embodiments in the present application and the features in the embodiments can be combined with each other without conflict.

[0039] ​In the description of the present application, it needs to be understood that the orientation or positional relationship indicated by the terms "center", "longitudinal", "transverse", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer" and the like is based on the orientation or positional relationship shown in the drawings, and is only for the convenience of describing the present application and simplifying the description, and does not indicate or imply that the device or element indicated must have a particular orientation, be constructed and operated in a particular orientation, and therefore cannot be understood as a limitation on the present application. In addition, the terms "first", "second" and the like are only for description purposes and cannot be understood as indicating or implying relative importance or implicitly indicating the number of the technical features indicated. Therefore, the features defined with "first", "second" and the like can be explicitly or implicitly included one or more. In the description of the present application, unless otherwise stated, the meaning of "a plurality of" is two or more.

[0040] In the description of the present application, it needs to be understood that unless otherwise explicitly specified and limited, the terms "mounting", "connecting", "connecting" should be understood broadly, for example, it can be fixedly connected, or it can be detachably connected, or integrally connected; it can be mechanically connected, or it can be electrically connected; it can be directly connected, or it can be indirectly connected through an intermediate medium, or it can be the communication inside two elements. For those skilled in the art, the specific meaning of the above terms in the present application can be understood through specific circumstances.

[0041] The present application will be described in detail below with reference to the accompanying drawings and in conjunction with embodiments.

[0042] As Figure 1 shown, the present application proposes an image sharpness evaluation method for long focal length visible light imaging system, specifically comprising the following steps:

[0043] S1: The focusing motor of the long focal length visible light imaging system drives the focusing lens to move to obtain a focusing image;

[0044] S2: The focusing image is preprocessed to obtain a preprocessed image, and the preprocessed image is sequentially subjected to global linear normalization and local contrast normalization processing to obtain a normalized image;

[0045] S3: The normalized image is subjected to NSST decomposition to obtain high-frequency sub-band coefficients of different scales and different directions;

[0046] S4: Based on the high-frequency sub-band coefficients of different scales and different directions, the direction consistency of different scales is calculated;

[0047] S5: Based on the direction consistency of different scales, a local sharpness function of different scales and different directions is constructed;

[0048] S6: the local sharpness values of different scales and different directions are weighted and fused to obtain a fused image;

[0049] S7: the pixel values corresponding to each pixel point of the fused image are averaged to obtain the sharpness evaluation value of the focus image.

[0050] It should be noted that the present application first adopts a median filter algorithm to pre-process the image, which is a typical representative of a non-linear filtering algorithm, which can eliminate isolated noise points and effectively protect edge contour information while removing random noise, which is very suitable for the high demand of edge information of the sharpness function. Then, the pre-processed image is decomposed by NSST (non-subsampled shearlet transform) to obtain low-frequency components and high-frequency subbands, and a robust image sharpness evaluation method is constructed by combining direction consistency and contrast normalization.

[0051] In addition, when the focus image is replaced, the replaced focus image is processed by S2-S7.

[0052] Further, NSST is a multi-scale geometric analysis tool based on shear wave theory, which has wide application in the field of image processing. Compared with traditional transform methods (such as wavelet transform, Contourlet transform, etc.), NSST has translation invariance, multi-scale, multi-direction and high efficiency of direction selection, etc. Its non-subsampled characteristics and precise filtering design ensure high accuracy of decomposition, which can better represent the geometric characteristics of the image, so the present application adopts NSST as the underlying framework for constructing the image sharpness evaluation function.

[0053] The present application pre-processes the image by median filtering, takes NSST (non-subsampled shearlet transform) as the basic framework, combines multi-scale decomposition and direction consistency, and can accurately evaluate the image sharpness of the long focal length visible light imaging system in complex scenes such as low contrast and noise interference. At the same time, through contrast normalization, the interference of exposure and contrast on the sharpness score is reduced, the score is more robust, the score range of different scenes is unified, and the consistency and comparability across scenes are improved.

[0054] In some embodiments, in step S2, the pre-processing operation is specifically a median filtering operation.

[0055] In some embodiments, in step S2, the calculation formula adopted for global linear normalization and local contrast normalization of the pre-processed image is:

[0056] ;

[0057] ;

[0058] wherein, is a pre-processed image, is a global linear normalized image, is a minimum value of pixels in the pre-processed image, is a maximum value of pixels in the pre-processed image, is an anti-zero term, is a coordinate position of a pixel point, is a local contrast normalized image, is a mean value of pixels in a local window W, is a standard deviation of pixels in the local window W.

[0059] It should be noted that any pixel point in the image to be processed is taken as the center to construct a local window with a radius of r, and the mean value and the standard deviation of the pixel gray scale in the local window are calculated, and the current pixel is normalized by using the mean value and the standard deviation to eliminate the influence of local brightness and contrast difference.

[0060] In some embodiments, in step S4, the calculation formula used for calculating the direction consistency of different scales is:

[0061] ;

[0062] wherein, is the direction consistency of the s scale, is the high-frequency sub-band coefficient of the s scale, is the s scale, and is the direction.

[0063] In some embodiments, the expression of the local sharpness function of different scales and different directions is:

[0064] ;

[0065] wherein, is the local sharpness value of the s scale, is the s scale, and is the direction. is the direction consistency of the s scale, is the high-frequency sub-band coefficient of the s scale, is the s scale, and is the direction.

[0066] In some embodiments, the calculation formula used for weighting and fusing the local sharpness values of different scales and different directions is:

[0067] ;

[0068] ;

[0069] ;

[0070] wherein, is the local sharpness value of the i-th scale and the j-th direction, is the weight coefficient of the i-th scale and the j-th direction, is the sub-band energy of the i-th scale and the j-th direction, is the sub-band energy of the i-th scale and the j-th direction, is the sub-band energy of the i-th scale and the j-th direction, is the sub-band energy of the i-th scale and the j-th direction, is the sub-band energy of the i-th scale and the j-th direction, is the sub-band energy of the i-th scale and the j-th direction, is the sub-band energy of the i-th scale and the j-th direction, is the fusion image, and N represents the number of pixels of the current sub-band.

[0071] Further, in different scenes and different distances, the contribution degree of each sub-band to the sharpness value of the whole image can be adjusted according to the richness degree of the high-frequency energy, that is, the value of is adjusted.

[0072] In some embodiments, the calculation formula for averaging the pixel values corresponding to each pixel point of the fusion image is:

[0073] ;

[0074] wherein, Score is the sharpness evaluation value.

[0075] It should be noted that after obtaining the sharpness evaluation value of the focus image, the sharpness evaluation value can be used to evaluate the image sharpness, which includes single image sharpness judgment, multiple image sharpness comparison, and the greater the sharpness evaluation value, the higher the overall image sharpness.

[0076] In a long focal length visible light imaging system, the sharpness evaluation value can be used as a feedback quantity to drive the focusing mechanism or control the adjustment of the imaging parameters. For example, when performing an automatic focusing process, the focusing motor drives the focusing lens to move within the focusing stroke, and a series of focusing images at different focusing positions are obtained. The obtained focusing images are evaluated by the present application to obtain the sharpness evaluation value, thereby assisting the long focal length visible light imaging system to form a closed loop control and improving the accuracy of automatic focusing.

[0077] It should be understood that the various forms of processes shown above can be reordered, added to, or deleted from. For example, the steps described in the present disclosure can be executed in parallel, in sequence, or in different orders, as long as the desired results of the technical solutions of the present disclosure can be achieved, which are not limited herein.

[0078] The above detailed description does not limit the scope of the application. Various modifications, combinations, sub-combinations and alternatives can be made to the detailed embodiment within the scope of the application. Any modification, equivalent replacement and improvement made without departing from the spirit and principle of the application shall fall within the scope of the application.

Claims

1. A method for evaluating image sharpness in long-focal-length visible light imaging systems, characterized in that: Specifically, the steps include the following: S1: The focusing motor of the long focal length visible light imaging system drives the focusing lens to move and obtain a focused image; S2: Perform preprocessing operations on the focusing image to obtain a preprocessed image. Then, perform global linear normalization and local contrast normalization on the preprocessed image in sequence to obtain a normalized image. In step S2, the calculation formulas used for global linear normalization and local contrast normalization of the preprocessed image are as follows: ； ； in, For preprocessing images, For globally linearly normalized images, The minimum pixel value in the preprocessed image. The maximum pixel value in the preprocessed image. To prevent zero items, The coordinates of the pixel. For locally contrast-normalized images, To obtain the average value of pixels within a local window W, To obtain the standard deviation of pixels within a local window W; S3: Perform NSST decomposition on the normalized image to obtain high-frequency subband coefficients at different scales and in different directions; S4: Calculate the directional consistency at different scales based on the high-frequency subband coefficients at different scales and in different directions; In step S4, the formula used to calculate the directional consistency at different scales is as follows: ； in, For directional consistency at the s-scale, For the first Scale, First High-frequency subband coefficients in each direction; S5: Based on the directional consistency at different scales, construct local sharpness functions for different scales and directions; The expressions for the local sharpness function at different scales and in different directions are: ； in, For the first Scale, First Local sharpness values ​​in each direction, For directional consistency at the s-th scale, For the first Scale, First High-frequency subband coefficients in each direction; S6: Weighted fusion of local sharpness values ​​at different scales and in different directions to obtain a fused image; S7: Take the average of the pixel values ​​corresponding to each pixel in the fused image to obtain the sharpness evaluation value of the focused image.

2. The image sharpness evaluation method for long focal length visible light imaging systems according to claim 1, characterized in that: In step S2, the preprocessing operation is specifically a median filtering operation.

3. The image sharpness evaluation method for long focal length visible light imaging systems according to claim 1, characterized in that: The calculation formula used for weighted fusion of local sharpness values ​​at different scales and in different directions is as follows: ； ； ； in, For the first Scale, First Local sharpness values ​​in each direction, For the first Scale, First Weight coefficients for each direction, For the first Scale, First Subband energy in each direction, To merge images.

4. The image sharpness evaluation method for long focal length visible light imaging systems according to claim 1, characterized in that: The formula used to average the pixel values ​​corresponding to each pixel in the fused image is as follows: ； Among them, the score is the clarity rating.

Images