Image processing method

Abstract

The embodiment of the invention discloses an image processing method, and the method comprises the steps: carrying out the polarization detection of a target object through a polarization detector, and obtaining a pixel response voltage corresponding to each pixel of the polarization detector; acquiring a polarization degree image and a polarization angle image according to the pixel response voltage corresponding to each pixel; according to the method, the infrared image corresponding to the target object is obtained, and the infrared image, the polarization degree image and the polarization angle image are fused to obtain the infrared polarization fusion image, so that the infrared image, the polarization degree image and the polarization angle image are fused to obtain the infrared polarization fusion image; according to the method, an infrared image, a polarization degree image and a polarization angle image are fused to obtain an infrared polarization fusion image, so that the infrared polarization fusion image contains polarization information and infrared information, and an infrared dim target and an infrared camouflage target can be effectively detected based on the polarization information in the infrared polarization fusion image subsequently.

Description

Technical Field

[0001] The present application relates to the technical field of image processing, and in particular to an image processing method. Background Art

[0002] Since infrared light was discovered by humans, it has been gradually applied to various industries and sectors such as industry, agriculture, medicine, and transportation, such as online power detection, underground mine temperature and gas measurement, mineral resource exploration, landform or environmental monitoring, weather forecasting, crop or environmental monitoring, etc., which has led to the rapid development of infrared detection technology.

[0003] After more than two centuries of development and research on infrared photoelectric detectors, the number of pixels in infrared detectors has grown from less than 100 yuan in the first generation to 100,000 yuan in the second generation, and then to millions of pixels in the third generation. After the emergence of a large family of infrared detector devices, people have entered the third generation with high sensitivity, large array, low cost and multi-spectrum as its technical characteristics.

[0004] With the rapid development of infrared technology, infrared camouflage technology is also developing rapidly. Technologies such as infrared camouflage paint and interference background have brought great challenges to traditional infrared detection technology. Infrared detection technology uses radiation intensity information to perform imaging. Therefore, a certain degree of radiation intensity difference is required when searching for a target to effectively detect the target. This makes it difficult to effectively detect infrared dim targets and infrared camouflaged targets in principle. Summary of the invention

[0005] In order to solve or partially solve the above problems, the present application provides an image processing method.

[0006] The present application proposes an image processing method, which includes: performing polarization detection on a target object through a polarization detector to obtain a pixel response voltage corresponding to each pixel of the polarization detector; obtaining a polarization degree image and a polarization angle image according to the pixel response voltage corresponding to each pixel; obtaining an infrared image corresponding to the target object, and fusing the infrared image, the polarization degree image and the polarization angle image to obtain an infrared polarization fused image.

[0007] Compared with the prior art, this application has the following beneficial effects:

[0008] The present application provides an image processing method. The method performs polarization detection on a target object through a polarization detector to obtain a pixel response voltage corresponding to each pixel of the polarization detector; obtains a polarization degree image and a polarization angle image according to the pixel response voltage corresponding to each pixel; obtains an infrared image corresponding to the target object, fuses the infrared image, the polarization degree image and the polarization angle image to obtain an infrared polarization fusion image. The method fuses the infrared image with the polarization degree image and the polarization angle image to obtain an infrared polarization fusion image, so that the infrared polarization fusion image contains polarization information and infrared information, so that infrared dim targets and infrared camouflaged targets can be effectively detected based on the polarization information in the infrared polarization fusion image, thereby avoiding the problem that a simple infrared image cannot effectively detect infrared dim targets and infrared camouflaged targets. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] Figure 1 It is a basic schematic diagram of an image processing method shown in the present application;

[0010] Figure 2 It is a basic schematic diagram of an M×N array polarization detector shown in this application. DETAILED DESCRIPTION

[0011] Here, exemplary embodiments will be described in detail, examples of which are shown in the accompanying drawings. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The implementations described in the following exemplary embodiments do not represent all implementations consistent with the present application. Instead, they are only examples of devices and methods consistent with some aspects of the present application as detailed in the attached claims.

[0012] The block diagrams shown in the accompanying drawings are merely functional entities and do not necessarily correspond to physically independent entities. That is, these functional entities may be implemented in software form, or in one or more hardware modules or integrated circuits, or in different networks and / or processor devices and / or microcontroller devices.

[0013] The flowcharts shown in the accompanying drawings are only exemplary and do not necessarily include all the contents and operations / steps, nor must they be executed in the order described. For example, some operations / steps can be decomposed, and some operations / steps can be combined or partially combined, so the actual execution order may change according to actual conditions.

[0014] It should also be noted that the "multiple" mentioned in this application refers to two or more than two. "And / or" describes the association relationship of the associated objects, indicating that there can be three relationships. For example, A and / or B can mean: A exists alone, A and B exist at the same time, and B exists alone. The character " / " generally indicates that the previous and next associated objects are in an "or" relationship.

[0015] Example

[0016] like Figure 1 As shown, Figure 1 An image processing method is proposed for this example, the method comprising:

[0017] S101, performing polarization detection on a target object by using a polarization detector to obtain a pixel response voltage corresponding to each pixel of the polarization detector;

[0018] S102, acquiring a polarization degree image and a polarization angle image according to a pixel response voltage corresponding to each pixel;

[0019] S103, acquiring an infrared image corresponding to the target object, and fusing the infrared image, the polarization degree image, and the polarization angle image to obtain an infrared polarization fused image.

[0020] Specifically, the present application performs polarization detection on the target object through a polarization detector with a polarization superpixel structure to obtain a pixel response voltage corresponding to each pixel of the polarization detector; illustratively, the polarization detector is an infrared polarization detector with a 2X2 polarization superpixel structure, such as Figure 2 As shown, the infrared polarization detector is an M×N array infrared polarization detector, and the polarization directions are 0°, 45°, 90° and 135° respectively.

[0021] In some examples, before performing polarization detection on a target object through a polarization detector, the present application also requires calibrating the polarization detector in a polarization direction, and the correction method includes but is not limited to a two-point correction or a multi-point correction. In view of the characteristics of the polarization detector, the present application first performs correction in the polarization direction, that is, correcting the pixels in the same polarization direction, rather than correcting the entire array together, so that the imaging in the polarization direction is more uniform and accurate.

[0022] In some examples, the pixel response voltage described in this application is the absolute value of the difference between the signal voltage output by the polarization detector to the target object and the shutdown level of the detector readout circuit. This application will obtain the polarization degree image and the polarization angle image based on the pixel response voltage corresponding to each pixel.

[0023] It can be understood that the present application will also acquire an infrared image corresponding to the target object, and the shooting angle of the infrared image is the same as the infrared image acquisition method of the polarization image. The infrared image can be acquired by an infrared polarization detector, and the infrared image can also be an infrared image obtained by another separately set infrared detector shooting the target object at the same angle as when the polarization detector shoots the polarization image.

[0024] After acquiring the infrared image, the present application will also fuse the infrared image, polarization degree image and polarization angle image to obtain an infrared polarization fused image. It can be understood that the present application does not limit the specific means of the infrared image, polarization degree image and polarization angle image.

[0025] For example, based on the pixel-level fusion of the RGB color space, the three channels are independent of each other and have the same weights. The above-mentioned infrared image, polarization degree image and polarization angle image are respectively mapped to the three channels of R, G, and B to complete the fusion and obtain an infrared polarization fusion image.

[0026] For example, based on the pixel-level fusion of the YCbCr color space, the three channels are independent of each other and have the same weights. The above infrared image, polarization degree image and polarization angle image are respectively mapped to the three channels of Y, Cb and Cr to complete the fusion and obtain the infrared polarization fusion image.

[0027] For another example, in the fusion algorithm based on the HSV color space, the three channels are independent of each other, the H channel has the highest weight, the S channel has the second highest weight, and the V channel has the lowest weight. The above polarization image, polarization angle image, and infrared image are respectively mapped to the three channels of H, S, and V to complete the fusion and obtain an infrared polarization fusion image. During the fusion algorithm operation, the weight of each image is distinguished, which amplifies the proportion of polarization information. According to different image processing requirements, the correspondence between the polarization image, polarization angle image, and infrared image and the three channels of H, S, and V can be adjusted to obtain an infrared polarization fusion image that highlights different information.

[0028] It can be understood that if the fused infrared polarization fusion image has the problem of unclear details, it is necessary to modulate and stretch at least one of the polarization image, polarization angle image, and infrared image; specifically, by setting an image modulation coefficient, multiplying it with the image to be stretched, and then fusing the resulting image with the other images. Generally, the modulation coefficient increases monotonically, but nonlinearly. The larger the value, the more the modulation coefficient needs to be increased. The specific form is not limited, so as to stretch the contrast between the target objects. Examples of modulation coefficients are as follows:

[0029] Example 1: P DoLP (x, y) = DoLP(x, y);

[0030] Example 2: P DoLP(x, y)=log[1+DoLP(x, y)];

[0031] Example 3: P Orient (x, y) = 1 + Orient (x, y).

[0032] The above P DoLP (x, y) is the modulation coefficient corresponding to the polarization image, DoLP(x, y) is the polarization image, POrient(x, y) is the modulation coefficient corresponding to the polarization angle image, Orient(x, y) is the polarization angle image, and (x, y) represents the pixel coordinates.

[0033] The image processing method provided in this example includes: performing polarization detection on the target object through a polarization detector to obtain a pixel response voltage corresponding to each pixel of the polarization detector; obtaining a polarization degree image and a polarization angle image according to the pixel response voltage corresponding to each pixel; obtaining an infrared image corresponding to the target object, fusing the infrared image, the polarization degree image and the polarization angle image to obtain an infrared polarization fused image. The method realizes the fusion of the infrared image with the polarization degree image and the polarization angle image to obtain an infrared polarization fused image. The method fuses the infrared image with the polarization degree image and the polarization angle image to obtain an infrared polarization fused image, so that the infrared polarization fused image contains polarization information and infrared information, so that the infrared dim target and infrared camouflaged target can be effectively detected based on the polarization information in the infrared polarization fused image in the subsequent process, thereby avoiding the problem that a simple infrared image cannot effectively detect an infrared dim target and an infrared camouflaged target.

[0034] In some examples, a polarization degree image and a polarization angle image are obtained based on a pixel response voltage corresponding to each pixel, including: calculating a first vector, a second vector, and a third vector based on the pixel response voltage corresponding to the pixel; calculating the polarization degree corresponding to the pixel based on the first vector, the second vector, and the third vector, and obtaining a polarization map image based on the polarization degree corresponding to each pixel; calculating the polarization angle corresponding to the pixel based on the second vector and the third vector, and obtaining a polarization angle image based on the polarization angle corresponding to each pixel.

[0035] Specifically, the present application inputs the pixel response voltage corresponding to the pixel into the Stokes formula to obtain the first vector, the second vector and the third vector; specifically as follows:

[0036]

[0037] S1=V S (0)-V S (90);

[0038] S2=V S (45)-V S (135).

[0039] Among them, the above V S (0) is the response voltage of the pixel with 0 degree polarization direction; V S (45) is the response voltage of the pixel with a polarization direction of 45 degrees, V S (90) is the response voltage of the pixel with a polarization direction of 90 degrees, V S (135) is the response voltage of the pixel with a polarization direction of 135 degrees, S0 is the first vector, S1 is the second vector, and S2 is the third vector.

[0040] The present application will also calculate the polarization degree corresponding to the pixel according to the first vector, the second vector and the third vector, and obtain a polarization image according to the polarization degree corresponding to each pixel; calculate the polarization angle corresponding to the pixel according to the second vector and the third vector, and obtain a polarization angle image according to the polarization angle corresponding to each pixel, as follows:

[0041]

[0042] The present application respectively substitutes the first vector, the second vector and the third vector into the above formula to obtain the corresponding polarization degree image and polarization angle image, thereby realizing the conversion of the 2X2 pixels of the superpixel into a single polarization pixel, and reducing the array size of the initial infrared polarization imaging image to half of the original size, that is, for an M×N infrared polarization image, a polarization image with an array size of (M / 2)×(N / 2) can be obtained. Instead, the polarization degree and polarization angle are calculated using any 2×2 adjacent 4 pixels, and for an M×N infrared image, a polarization image with an array size of (M-1)×(N-1) can be obtained, which significantly increases the array size of the polarization image.

[0043] In some examples, in order to stretch the value ranges of the three images, namely, the infrared image, the polarization image, and the polarization angle image, taking into account their respective characteristics, and obtain the infrared image, the polarization image, and the polarization angle image with the same or similar value ranges, the present application will obtain the first range, the second range, and the third range corresponding to the infrared image, the polarization image, and the polarization angle image, respectively; determine the stretching coefficients corresponding to the infrared image, the polarization image, and the polarization angle image, respectively, according to the first range, the second range, and the third range; stretch the infrared image, the polarization image, and the polarization angle image according to the stretching coefficient corresponding to each image, and fuse the stretched infrared image, the polarization image, and the polarization angle image to obtain an infrared polarization fused image.

[0044] Specifically, the stretching coefficient corresponding to the infrared image is multiplied with the infrared image, the stretching coefficient corresponding to the polarization image is multiplied with the polarization image, and the stretching coefficient corresponding to the polarization angle image is multiplied with the polarization angle image. Finally, the three images after stretching are fused to obtain an infrared polarization fused image that highlights different information. The above method stretches the infrared image, the polarization degree image and the polarization angle image to obtain an infrared image, a polarization degree image and a polarization angle image with the same or similar value range, and then obtains an infrared polarization fused image that highlights different information.

[0045] In some examples, the stretching coefficients corresponding to the infrared image, the polarization image, and the polarization angle image are determined according to the first range, the second range, and the third range, including: multiplying the second range by the third range to obtain the stretching coefficient corresponding to the infrared image; multiplying the first range by the third range to obtain the stretching coefficient corresponding to the polarization image; multiplying the second range by the first range to obtain the stretching coefficient corresponding to the polarization angle image.

[0046] The first measurement range corresponding to the infrared image is A1, the second measurement range corresponding to the polarization image is A2, and the third measurement range corresponding to the polarization angle image is A3. The stretching coefficient corresponding to the infrared image is recorded as i1, the stretching coefficient corresponding to the polarization image is recorded as i2, and the stretching coefficient corresponding to the polarization angle image is recorded as i3.

[0047] i1=A2×A3;

[0048] i2=A1×A3;

[0049] i3=A2×A1.

[0050] Then, i1 is multiplied with the infrared image, i2 is multiplied with the polarization image, and i3 is multiplied with the polarization angle image. Finally, the three processed images are fused to obtain an infrared polarization fused image that highlights different information. The above method takes into account the ranges of the infrared image, the polarization image, and the polarization angle image. The value ranges of the three images, namely, the infrared image, the polarization image, and the polarization angle image, are stretched by taking into account their respective characteristics through the range corresponding to each image, so as to obtain infrared images, polarization images, and polarization angle images with the same or similar value ranges, and then obtain an infrared polarization fused image that highlights different information.

[0051] In some examples, the stretching coefficients corresponding to the infrared image, the polarization image, and the polarization angle image, respectively, are determined according to the first range, the second range, and the third range, including: obtaining the first arithmetic mean, the second arithmetic mean, and the third arithmetic mean corresponding to the infrared image, the polarization image, and the polarization angle image, respectively; calculating the range sum of the first range, the second range, and the third range, and calculating the arithmetic mean sum of the first arithmetic mean, the second arithmetic mean, and the third arithmetic mean; calculating the stretching coefficient corresponding to the infrared image according to the range sum, the arithmetic mean sum, the first range, and the first arithmetic mean; calculating the stretching coefficient corresponding to the polarization image according to the range sum, the arithmetic mean sum, the second range, and the second arithmetic mean; calculating the stretching coefficient corresponding to the polarization angle image according to the range sum, the arithmetic mean sum, the third range, and the third arithmetic mean.

[0052] The first measurement range corresponding to the infrared image is A1, the second measurement range corresponding to the polarization image is A2, and the third measurement range corresponding to the polarization angle image is A3. The first arithmetic mean corresponding to the infrared image is B1, the second arithmetic mean corresponding to the polarization image is B2, and the third arithmetic mean corresponding to the polarization angle image is B3. The stretching coefficient corresponding to the infrared image is recorded as i1, the stretching coefficient corresponding to the polarization image is recorded as i2, and the stretching coefficient corresponding to the polarization angle image is recorded as i3.

[0053]

[0054] Among them, the above x is a preset weighting coefficient used to adjust the proportion of the range and the arithmetic mean, and the value range of x is: 0<x≤2. It can be understood that the arithmetic mean is the expectation of the array in probability theory and can better represent the situation of the array. If the range and the arithmetic mean are taken into account, the influence of a very small amount of extreme values ​​on the stretching distortion of the array can be eliminated. The range and arithmetic mean of the infrared image, polarization image and polarization angle image are taken into account by the above method. The range of the three images of the infrared image, polarization image and polarization angle image is stretched by taking into account their respective characteristics through the range and arithmetic mean corresponding to each image, and the infrared image, polarization image and polarization angle image with the same or similar value range are obtained, and then the infrared polarization fusion image highlighting different information is obtained, which improves the accuracy of the infrared polarization fusion image.

[0055] In some examples, the stretching coefficients corresponding to the infrared image, the polarization degree image, and the polarization angle image, respectively, are determined according to the first range, the second range, and the third range, including: obtaining the first median, the second median, and the third median corresponding to the infrared image, the polarization degree image, and the polarization angle image, respectively; calculating the range sum of the first range, the second range, and the third range, and calculating the median sum of the first median, the second median, and the third median; calculating the stretching coefficient corresponding to the infrared image according to the range sum, the median sum, the first range, and the first median; calculating the stretching coefficient corresponding to the polarization degree image according to the range sum, the median sum, the second range, and the second median; calculating the stretching coefficient corresponding to the polarization angle image according to the range sum, the median sum, the third range, and the third median.

[0056] The first measurement range corresponding to the infrared image is A1, the second measurement range corresponding to the polarization image is A2, and the third measurement range corresponding to the polarization angle image is A3. The first median corresponding to the infrared image is C1, the second median corresponding to the polarization image is C2, and the third median corresponding to the polarization angle image is C3. The stretch coefficient corresponding to the infrared image is recorded as i1, the stretch coefficient corresponding to the polarization image is recorded as i2, and the stretch coefficient corresponding to the polarization angle image is recorded as i3.

[0057]

[0058]

[0059] Among them, the above x is a preset weighting coefficient used to adjust the proportion of the range and the median, and the value range of x is: 0<x≤2. It can be understood that the median can represent the situation of the array. If the range and the median are taken into account, the influence of a very small amount of the maximum value on the array stretching distortion can be eliminated. The above method takes into account the range and median of the infrared image, polarization image and polarization angle image. The range of the three images of the infrared image, polarization image and polarization angle image is stretched by the range and median corresponding to each image, taking into account their respective characteristics, and obtaining infrared images and polarization images and polarization angle images with the same or similar value ranges, and then obtaining an infrared polarization fusion image that highlights different information, thereby improving the accuracy of the infrared polarization fusion image.

[0060] In some examples, the stretching coefficients corresponding to the infrared image, the polarization image, and the polarization angle image, respectively, are determined according to the first range, the second range, and the third range, including: obtaining the first median, the second median, and the third median corresponding to the infrared image, the polarization image, and the polarization angle image, respectively; obtaining the first arithmetic mean, the second arithmetic mean, and the third arithmetic mean corresponding to the infrared image, the polarization image, and the polarization angle image, respectively; calculating the range sum of the first range, the second range, and the third range, calculating the median sum of the first median, the second median, and the third median, and calculating the arithmetic mean sum of the first arithmetic mean, the second arithmetic mean, and the third arithmetic mean; calculating the stretching coefficient corresponding to the infrared image according to the range sum, the median sum, the arithmetic mean sum, the first range, the first arithmetic mean, and the first median; calculating the stretching coefficient corresponding to the polarization image according to the range sum, the median sum, the arithmetic mean sum, the second range, the second arithmetic mean, and the second median; calculating the stretching coefficient corresponding to the polarization angle image according to the range sum, the median sum, the arithmetic mean sum, the third range, the third arithmetic mean, and the third median.

[0061] Among them, let the first range corresponding to the infrared image be A1, the second range corresponding to the polarization image be A2, and the third range corresponding to the polarization angle image be A3. Let the first arithmetic mean corresponding to the infrared image be B1, the second arithmetic mean corresponding to the polarization image be B2, and the third arithmetic mean corresponding to the polarization angle image be B3. Let the first median corresponding to the infrared image be C1, the second median corresponding to the polarization image be C2, and the third median corresponding to the polarization angle image be C3. Let the stretch coefficient corresponding to the infrared image be i1, the stretch coefficient corresponding to the polarization image be i2, and the stretch coefficient corresponding to the polarization angle image be i3. Among them:

[0062]

[0063] Among them, the above x and y are pre-set weighting coefficients used to adjust the proportion of range, median and arithmetic mean, and the value range of x is: 0<x≤2; the value range of y is: 0<y≤2. It can be understood that the arithmetic mean and median can represent the situation of the array. If the range, median and average value are taken into account, the influence of a very small amount of extreme values ​​on the stretching distortion of the array can be eliminated. The range, median and arithmetic mean of the infrared image, polarization image and polarization angle image are taken into account by the above method. The range of the three images of infrared image, polarization image and polarization angle image is stretched by taking into account their respective characteristics through the range, median and arithmetic mean corresponding to each image, and the infrared image, polarization image and polarization angle image with the same or similar value range are obtained, and then the infrared polarization fusion image highlighting different information is obtained, which improves the accuracy of the infrared polarization fusion image.

[0064] In order to better understand the present application, this embodiment provides a more specific example for explanation:

[0065] This example selects an infrared integrated polarization detector for fusion imaging (model but not limited to MW640X512), the polarization superpixel is 2X2, and the polarization directions are 0°, 45°, 90° and 135° respectively.

[0066] Two-point correction is performed in each polarization direction, and then the target is focused and imaged to collect the correction data of each pixel signal voltage. Then the difference is made with the shutdown level of the detector readout circuit to obtain the response voltage Vs in the Stokes formula. Taking any 2×2 adjacent 4 pixels as a unit, according to the Stokes vector method, the linear polarization degree image and polarization angle image of the 639×511 array scale are calculated.

[0067] Stokes vector calculation method:

[0068]

[0069] S1=V S (0)-V S (90);

[0070] S2=V S (45)-V S (135).

[0071] Polarization degree and polarization angle calculation method:

[0072]

[0073] The 3X3 weighted median filter method is used to perform blind pixel replacement on the infrared image and the polarization degree image and polarization angle image obtained in the first step to reduce noise. The weighting method is

[0074] The value ranges of the infrared image, polarization degree image and polarization angle image are stretched taking into account their respective characteristics. The ranges of the infrared image, polarization degree image and polarization angle image are A1, A2 and A3 respectively, and their arithmetic means are B1, B2 and B3 respectively.

[0075] Taking into account the range and arithmetic mean, the infrared image, polarization degree image and polarization angle image are multiplied by the following coefficients respectively:

[0076]

[0077] In the above formula, i1, i2 and i3 are the stretching coefficients of the infrared image, polarization degree image and polarization angle image respectively. The stretched image can be obtained by multiplying the stretching coefficient by the corresponding image.

[0078] The fusion algorithm based on HSV color space is adopted to correspond the polarization degree image, polarization angle image and infrared image to the three channels of H, S and V respectively, and then the fusion can be completed.

[0079] There is a problem that the details in the fused image are not clear enough. Set an image modulation coefficient, multiply it with the image to be stretched, and then fuse it. The modulation coefficient is as follows:

[0080] P DoLP (x, y)=log[1+DoLP(x, y)].

[0081] In the above formula, P DoLP (x, y) is a monotonically continuous modulation coefficient, DoLP(x, y) is a polarization image, and (x, y) represents pixel coordinates.

[0082] After the above image processing, a fused imaging image with both infrared information and polarization information is obtained.

[0083] Another example

[0084] An infrared integrated polarization detector is selected for fusion imaging (the models of the infrared integrated polarization detector include but are not limited to: LW320X256), the polarization superpixel is 2X2 in specification, and the polarization directions are 0°, 45°, 90° and 135° respectively.

[0085] Two-point correction is performed in each polarization direction, and then the target is focused and imaged to collect the correction data of each pixel signal voltage. Then the difference is made with the shutdown level of the detector readout circuit to obtain the response voltage Vs in the Stokes formula. Taking any 2×2 adjacent 4 pixels as a unit, according to the Stokes vector method, the linear polarization degree image and polarization angle image of the 319×255 array scale are calculated.

[0086] Stokes vector calculation method:

[0087]

[0088] S1=V S (0)-V S (90);

[0089] S2=V S (45)-V S (135).

[0090] Calculation method of linear polarization degree and polarization angle:

[0091]

[0092] The 3X3 median filter method is used to perform blind pixel replacement on the infrared image and the polarization degree image and polarization angle image obtained in the first step to reduce noise. The weighting method is

[0093] The value ranges of the infrared image, polarization degree image and polarization angle image are stretched taking into account their respective characteristics. The ranges of the infrared image, polarization degree image and polarization angle image are A1, A2 and A3 respectively, and their arithmetic means are B1, B2 and B3 respectively.

[0094] Taking into account the range and arithmetic mean, the infrared image, polarization degree image and polarization angle image are multiplied by the following coefficients respectively:

[0095]

[0096] In the above formula, i1, i2 and i3 are the stretching coefficients of the infrared image, polarization degree image and polarization angle image respectively. The stretched image can be obtained by multiplying the stretching coefficient by the corresponding image.

[0097] The fusion algorithm based on HSV color space is adopted to correspond the polarization angle image, polarization degree image and infrared image to the three channels of H, S and V respectively, and then the fusion can be completed.

[0098] There is a problem that the details in the fused image are not clear enough. Set an image modulation coefficient, multiply it with the image to be stretched, and then fuse it. The modulation coefficient is as follows:

[0099] P Orient (x, y) = 1 + Orient (x, y).

[0100] In the above formula, POrient(x, y) is the polarization phase angle modulation coefficient, Orient(x, y) is the polarization angle image, and (x, y) represents the pixel coordinates. After the above image processing, a fused image with both infrared information and polarization information is obtained.

[0101] The above scheme of the present application is aimed at the characteristics of multiple polarization direction detectors. First, the polarization direction is corrected, that is, the pixels in the same polarization direction are corrected, not the entire array, so that the polarization direction imaging is more uniform and accurate, and the present application innovatively defines the response voltage in the Stokes formula. Then the Stokes formula is used to calculate the vector component values, and then the corresponding polarization degree image and polarization angle image are obtained through the formulas of polarization degree and polarization angle. The present application innovatively proposes to calculate polarization data with any 2×2 adjacent 4 pixels, instead of the traditional calculation of polarization data with each 2×2 super pixel. While making full use of polarization information without distortion, it greatly improves the array scale of polarization data and improves the spatial resolution of the final polarization fusion image.

[0102] This application first performs blind pixel replacement in a conventional way, i.e. median filtering or mean filtering or median weighted filtering or mean weighted filtering, and then innovatively proposes to stretch the infrared image, polarization degree image and polarization angle image to eliminate the fusion distortion caused by the large range difference. Generally, the range of polarization degree and polarization angle is much smaller than that of infrared image. Without stretching and direct fusion, the polarization information will be seriously lost, making it impossible to display the polarization characteristics. This patent takes into account the range, arithmetic mean and median of the image, and provides a variety of stretching coefficient methods that can be used in practice, which scientifically increases the proportion of polarization information.

[0103] This application adopts a fusion algorithm based on color space, and gives the characteristics and advantages of various color space fusion algorithms based on different purposes. In terms of highlighting polarization information, the fusion algorithm based on HSV color space is preferred, which can highlight different image information in a targeted manner. In addition, on the basis of the fusion algorithm, the modulation coefficient is added to compensate for the shortcomings of the fusion algorithm, further improve the contrast between scenes, and stretch the details of the image.

[0104] After the above image processing, an infrared polarization fusion image with both infrared information and polarization information can be obtained, which greatly enhances the contribution of polarization characteristic information to detection and imaging.

[0105] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic, for example, the division of units is only a logical function division, and there may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed.

[0106] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the embodiments of the present invention.

[0107] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0108] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods of each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program codes.

[0109] The technical solutions provided by the embodiments of the present invention are introduced in detail above. Specific examples are used in this patent to illustrate the principles and implementation methods of the embodiments of the present invention. The description of the above embodiments is only applicable to help understand the principles of the embodiments of the present invention. The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. An image processing method, characterized in that: The method comprises: Performing polarization detection on the target object by using a polarization detector to obtain a pixel response voltage corresponding to each pixel of the polarization detector; Acquire a polarization degree image and a polarization angle image according to the pixel response voltage corresponding to each pixel; An infrared image corresponding to the target object is acquired, and the infrared image, the polarization degree image, and the polarization angle image are fused to obtain an infrared polarization fused image.

2. The method according to claim 1, characterized in that Acquiring a polarization degree image and a polarization angle image according to the pixel response voltage corresponding to each pixel, including: Calculating a first vector, a second vector and a third vector according to the pixel response voltage corresponding to the pixel; Calculate the polarization degree corresponding to the pixel according to the first vector, the second vector and the third vector, and obtain the polarization map image according to the polarization degree corresponding to each pixel; The polarization angle corresponding to the pixel is calculated according to the second vector and the third vector, and the polarization angle image is obtained according to the polarization angle corresponding to each pixel.

3. The method according to claim 1, characterized in that The infrared image, the polarization degree image and the polarization angle image are fused to obtain an infrared polarization fused image, including: Acquire a first measurement range, a second measurement range, and a third measurement range respectively corresponding to the infrared image, the polarization degree image, and the polarization angle image; Determine, according to the first measurement range, the second measurement range, and the third measurement range, stretch coefficients corresponding to the infrared image, the polarization degree image, and the polarization angle image, respectively; The infrared image, the polarization degree image and the polarization angle image are stretched according to the stretching coefficient corresponding to each image, and the stretched infrared image, the polarization degree image and the polarization angle image are fused to obtain the infrared polarization fused image.

4. The method according to claim 1, characterized in that Determining the stretching coefficients corresponding to the infrared image, the polarization degree image, and the polarization angle image respectively according to the first measurement range, the second measurement range, and the third measurement range includes: Multiplying the second measurement range by the third measurement range to obtain a stretching coefficient corresponding to the infrared image; Multiplying the first measurement range and the third measurement range to obtain a stretch coefficient corresponding to the polarization degree image; The second measurement range is multiplied by the first measurement range to obtain a stretch coefficient corresponding to the polarization angle image.

5. The method according to claim 1, characterized in that Determining the stretching coefficients corresponding to the infrared image, the polarization degree image, and the polarization angle image respectively according to the first measurement range, the second measurement range, and the third measurement range includes: Obtaining a first arithmetic mean value, a second arithmetic mean value, and a third arithmetic mean value corresponding to the infrared image, the polarization degree image, and the polarization angle image, respectively; Calculating the sum of the first measurement range, the second measurement range, and the third measurement range, and calculating the sum of the arithmetic mean of the first arithmetic mean, the second arithmetic mean, and the third arithmetic mean; Calculating a stretch coefficient corresponding to the infrared image according to the sum of the measurement ranges, the sum of the arithmetic mean values, the first measurement range, and the first arithmetic mean value; Calculating a stretch coefficient corresponding to the polarization degree image according to the sum of the measurement range, the sum of the arithmetic mean, the second measurement range, and the second arithmetic mean; The stretching coefficient corresponding to the polarization angle image is calculated according to the sum of the ranges, the sum of the arithmetic mean values, the third range and the third arithmetic mean value.

6. The method according to claim 1, characterized in that Determining the stretching coefficients corresponding to the infrared image, the polarization degree image, and the polarization angle image respectively according to the first measurement range, the second measurement range, and the third measurement range includes: Obtaining a first median, a second median, and a third median corresponding to the infrared image, the polarization degree image, and the polarization angle image, respectively; Calculate the sum of the first range, the second range, and the third range, and calculate the sum of the medians of the first median, the second median, and the third median; Calculating a stretch coefficient corresponding to the infrared image according to the sum of the ranges, the sum of the medians, the first range, and the first median; Calculating a stretch coefficient corresponding to the polarization degree image according to the range sum, the median sum, the second range, and the second median; The stretching coefficient corresponding to the polarization angle image is calculated according to the range sum, the median sum, the third range and the third median.

7. The method according to claim 1, characterized in that Determining the stretching coefficients corresponding to the infrared image, the polarization degree image, and the polarization angle image respectively according to the first measurement range, the second measurement range, and the third measurement range includes: Obtaining a first median, a second median, and a third median respectively corresponding to the infrared image, the polarization degree image, and the polarization angle image; obtaining a first arithmetic mean, a second arithmetic mean, and a third arithmetic mean respectively corresponding to the infrared image, the polarization degree image, and the polarization angle image; Calculate the range sum of the first range, the second range and the third range, calculate the median sum of the first median, the second median and the third median, and calculate the arithmetic mean sum of the first arithmetic mean, the second arithmetic mean and the third arithmetic mean; Calculate the stretch coefficient corresponding to the infrared image according to the sum of the ranges, the sum of the medians, the sum of the arithmetic mean values, the first range, the first arithmetic mean value, and the first median value; Calculate the stretch coefficient corresponding to the polarization degree image according to the range sum, the median sum, the arithmetic mean sum, the second range, the second arithmetic mean and the second median; The stretching coefficient corresponding to the polarization angle image is calculated according to the range sum, the median sum, the arithmetic mean sum, the third range, the third arithmetic mean and the third median.

Images