Background noise suppression method for high-sensitivity images of space targets

Abstract

The present invention relates to the field of image denoising, and in particular to a method for suppressing background noise of a high-sensitivity image of a space target, comprising S1: tracking and imaging the space target using a light measuring device to obtain a continuous high-sensitivity image sequence; S2: counting the grayscale values ​​in the row and column directions in the high-sensitivity image sequence to obtain a brightness distribution; S3: determining the coordinates of the space target in the high-sensitivity image sequence based on the brightness distribution; S4: adjusting the exposure of the high-sensitivity camera in the light measuring device, re-imaging the space target, and obtaining a 16-bit original image, so that the brightness of the pixel points around the space target in the original image reaches a brightness threshold; S5: simultaneously selecting the high 8-bit data of the 16-bit original image and stretching the grayscale at equal intervals to obtain two 8-bit images; S6: fusing the two 8-bit images to obtain a denoised high-sensitivity image. The present invention effectively removes the noise of the high-sensitivity image of the space target.

Description

Technical Field

[0001] The present invention relates to the technical field of image denoising, and in particular to a method for suppressing background noise of a high-sensitivity image of a space target. Background Art

[0002] For high-sensitivity imaging of space targets, due to the long observation distance of space targets, they are often affected by factors such as atmospheric interference during imaging, resulting in weak imaging intensity and small imaging size of space targets. In high-sensitivity images, they mostly appear as tiny light spots, which are not significantly different in morphology from the point noise of high-sensitivity cameras. This adds a lot of difficulties to the denoising of high-sensitivity images of space targets.

[0003] There are many studies on image denoising methods in the field of optical imaging, but most algorithms focus on denoising of daily or life images. However, in high-sensitivity images of space targets, the images formed by space targets and noise are very different in morphology. The current traditional denoising methods will remove tiny space targets as noise. In addition, the working frame rate of optical measurement equipment is high, which places extremely high requirements on the real-time performance of image denoising methods, which also poses a great challenge to image denoising methods.

[0004] Traditional isotropic denoising methods mainly process the entire image uniformly through median filtering, mean filtering and other means. This method has the advantage of good real-time performance, but its disadvantages are also obvious. Since this type of denoising method will process all pixels in the image indiscriminately, the processed image will lose image details, resulting in reduced image sharpness and blurring. For high-sensitivity images of space targets, since the imaging size of space targets in the image is only a few dozen pixels, it is almost the same as the noise signal. After being processed by this type of denoising method, it will be directly filtered out as noise.

[0005] Traditional anisotropic denoising methods, such as guided filtering, often first determine whether the high-frequency signals in the image belong to the image contour or noise, and then remove these high-frequency signals in a targeted manner based on the judgment results. Such methods can avoid the reduction of image sharpness and blurring caused by filtering the image contour area. However, in high-sensitivity images of space targets, the space target imaging is only a point without obvious contours, and the imaging effect of the space target is not much different from the background noise, making it difficult to determine the contour of the space target. Moreover, such methods are time-consuming and cannot meet the task requirements of background noise suppression in high-sensitivity images of space targets.

[0006] At present, with the development of deep learning technology, various deep learning denoising models have been established, and many deep learning denoising methods have also achieved good results. However, there is still no data set related to denoising of high-sensitivity imaging of space targets. Since the space targets and noise in the high-sensitivity images of space targets are highly similar, it is even impossible to manually distinguish the space target area and the noise area in the image when annotating the data set. In addition, the deep learning denoising method is often time-consuming, so it is not suitable for background noise suppression of high-sensitivity images of space targets of optical measurement equipment with high real-time requirements. Summary of the invention

[0007] In view of this, the present invention aims to provide a method for suppressing background noise of high-sensitivity images of space targets, so as to solve the technical problem that traditional image denoising methods are not applicable to background noise suppression of high-sensitivity images of space targets.

[0008] To achieve the above object, the technical solution created by the present invention is implemented as follows:

[0009] A method for suppressing background noise of a high-sensitivity image of a space target comprises the following steps:

[0010] S1: Use optical measurement equipment to track and image space targets to obtain continuous high-sensitivity image sequences;

[0011] S2: Count the grayscale values ​​in the row and column directions in the high-sensitivity image sequence to obtain the brightness distribution;

[0012] S3: Determine the coordinates of the space target in the high-sensitivity image sequence based on the brightness distribution;

[0013] S4: adjusting the exposure of the high-sensitivity camera in the optical measurement device, re-imaging the space target, obtaining a 16-bit original image, and making the brightness of the pixel points around the space target in the 16-bit original image reach the brightness threshold;

[0014] S5: simultaneously select the high 8-bit data of the 16-bit original image and stretch the grayscale with equal spacing to obtain two 8-bit images;

[0015] S6: Fuse the two 8-bit images to obtain a denoised high-sensitivity image.

[0016] Further, it is assumed that the high-sensitivity image sequence includes n frames of high-sensitivity images, the width and height of each frame of the high-sensitivity image are W and H respectively, and (x, y) is the horizontal coordinate and vertical coordinate of the pixel point in the high-sensitivity image;

[0017] In step S2, the process of counting the grayscale values ​​in the row direction and the column direction in the high-sensitivity image sequence is as follows:

[0018] Sum the grayscale values ​​of each row of the i-th frame high-sensitivity image to obtain the row grayscale value Img_i_line(y), Img_i_line(y)=Img_i(0,y)+Img_i(1,y)+…+ Img_i(W-1,y), Img_i(0,y), Img_i(1,y), …, Img_i(W-1,y) respectively represent the grayscale values ​​of pixel (0,y), pixel (1,y), …, pixel (W-1,y) in the i-th frame high-sensitivity image;

[0019] Sum the grayscale values ​​of each column of the i-th frame of high-sensitivity image to obtain the column grayscale value Img_i_column(x), Img_i_column(x)=Img_i(x,0)+Img_i(x,1)+…+Img_i(x,H-1), where Img_i(x,0), Img_i(x,1), …, Img_i(x,H-1) represent the grayscale values ​​of pixel (x,0), pixel (x,1), …, pixel (x,H-1) in the i-th frame of high-sensitivity image, respectively;

[0020] Sum the grayscale value of each line of the n-frame high-sensitivity image to obtain the total grayscale value Img_line(y), Img_line(y)=Img_1_line(y)+Img_2_line(y)+…+Img_n_line(y), Img_1_line(y), Img_2_line(y), …, Img_n_line(y) represent the grayscale value of each line of the 1st to the nth frame high-sensitivity images respectively;

[0021] The column grayscale values ​​of each column of n frames of high-sensitivity images are summed to obtain the total column grayscale value Img_column(x), Img_column(x)=Img_1_column(x)+Img_2_column(x)+…+Img_n_column(x), Img_1_column(x), Img_2_column(x), …, Img_n_column(x) represent the column grayscale values ​​of each column of the 1st to nth frames of high-sensitivity images, respectively.

[0022] Furthermore, let the coordinates of the space target in the high-sensitivity image be (xn,yn);

[0023] In step S3, the process of determining the coordinates of the space target in the high-sensitivity image sequence based on the brightness distribution is as follows:

[0024] According to the gray value, the top 0.5% of the pixels in the total gray value of the column Img_column(x) and the total gray value of the row Img_line(y) are counted, and the discontinuous pixels in the statistical results are removed to obtain the abscissa set {x1,x2,…xa} and the ordinate set {y1,y2,…yb}, then:

[0025] xn=(x1+x2+…+xa) / a;

[0026] yn=(y1+y2+…+yb) / b.

[0027] Further, in step S4, the brightness threshold is set to 50000.

[0028] Furthermore, assuming that the 16-bit original image is Img0, and the grayscale value of the pixel point (x, y) in the original image Img0 is Img0(x, y), then in step S5, the high 8-bit data of the original image Img0 is selected, and the obtained 8-bit image is Img1, and the grayscale value Img0(x, y) of the pixel point (x, y) in the 8-bit image Img1 is converted to Img1(x, y) by the following formula:

[0029]

[0030] Furthermore, in step S5, the process of stretching the original image Img0 with equal grayscale spacing is as follows:

[0031] The grayscale values ​​of all pixels in the original image Img0 are counted to obtain a set of pixels in the grayscale value range of 0-65535;

[0032] Count the number of valid pixels in the pixel set, record it as count, and record it in ascending order;

[0033] The original image Img0 is transformed to obtain an 8-bit image Img2. The grayscale value Img0(x,y) of the pixel (x,y) in the 8-bit image Img2 is converted to Img2(x,y) by the following formula:

[0034]

[0035] Among them, t is the ranking of the gray value Img0(x,y) of the pixel (x,y) in the original image Img0 in the valid gray values ​​in ascending order.

[0036] Furthermore, in step S6, the two 8-bit images are fused to obtain a denoised high-sensitivity image Img3. The grayscale value of the high-sensitivity image Img3 is Img3(x, y), and the calculation formula is:

[0037]

[0038] Among them, μ is the fusion parameter.

[0039] Furthermore, after step S6, the following steps are also included:

[0040] S7: Use a stretching method to stretch the high-sensitivity image Img3.

[0041] Furthermore, the stretching method is a linear stretching method or a nonlinear stretching method.

[0042] Compared with the prior art, the present invention can achieve the following beneficial effects:

[0043] (1) The noise suppression speed is fast, taking only milliseconds to achieve the high real-time requirements of optical measurement equipment.

[0044] (2) Based on the fact that space target imaging is highly sensitive to camera dimming, while noise is less sensitive to camera dimming, this method can effectively solve the problem of difficulty in distinguishing between space targets and noise due to their similar morphology, thereby avoiding the situation where tiny space targets are treated as noise and removed together with the noise.

[0045] (3) It can appropriately protect the background while suppressing noise, and ensure that faint targets such as stars are not removed after imaging. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] The drawings constituting part of the present invention are used to provide a further understanding of the present invention. The exemplary embodiments and descriptions of the present invention are used to explain the present invention and do not constitute an improper limitation on the present invention. In the drawings:

[0047] Figure 1 A schematic flow chart of a method for suppressing background noise of a space target high-sensitivity image according to an embodiment of the present invention;

[0048] Figure 2 A schematic diagram of a nonlinear processing function according to an embodiment of the present invention;

[0049] Figure 3 A schematic diagram of a high-sensitivity image before denoising according to an embodiment of the present invention;

[0050] Figure 4 A schematic diagram of a high-sensitivity image after denoising according to an embodiment of the present invention. DETAILED DESCRIPTION

[0051] In order to make the purpose, technical solution and advantages of the invention more clear, the invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described here are only used to explain the invention and do not constitute a limitation of the invention.

[0052] It should be noted that, in the absence of conflict, the embodiments of the present invention and the features in the embodiments may be combined with each other.

[0053] In the description of the present invention, it should be understood that the terms "center", "longitudinal", "lateral", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside" and the like indicate positions or positional relationships based on the positions or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present invention. In addition, the terms "first", "second", etc. are only used for descriptive purposes, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, features defined as "first", "second", etc. may explicitly or implicitly include one or more of the features. In the description of the present invention, unless otherwise specified, "multiple" means two or more.

[0054] In the description of the invention, it should be noted that, unless otherwise clearly specified and limited, the terms "installation", "connection" and "connection" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the invention can be understood according to specific circumstances.

[0055] Since the images formed by the space target and the noise in the high-sensitivity image of the space target are very different in morphology, the tiny space target is easily filtered out as noise. In order to distinguish the space target and the noise in the high-sensitivity image, the present invention realizes the distinction between the space target and the noise based on the different dimming response characteristics of the space target and the noise to the high-sensitivity camera of the optical measuring device. The image formed by the space target is highly sensitive to the camera dimming, while the image formed by the noise is less sensitive to the camera dimming. Therefore, when the high-sensitivity camera is dimmed, the image formed by the space target can be adjusted to a brightness close to overexposure, that is, for a 16-bit high-sensitivity image, the grayscale value of the image formed by the space target can be adjusted to close to 60000, while the brightness of the noise at this time is low and often cannot reach a higher brightness. When the image formed by the space target is close to overexposure, the grayscale value of the noise is usually less than 7000, while the grayscale value of the background part is lower, usually not exceeding 2000. The present invention utilizes this prior knowledge to achieve noise removal.

[0056] The following will refer to Figure 1~Figure 4 The invention is described in detail with reference to the embodiments.

[0057] like Figure 1 As shown, an embodiment of the present invention provides a method for suppressing background noise of a high-sensitivity image of a space target, comprising the following steps:

[0058] S1: Use optical measurement equipment to track and image space targets to obtain continuous high-sensitivity image sequences.

[0059] For space target observation tasks, the motion trajectory of the space target is known. During the execution of the space target observation task, the optical measurement equipment will be controlled to track the space target according to the motion trajectory of the space target. The space target will be in the optical measurement field of view at this time, and will be in a relatively stable position in the high-sensitivity image. It can be considered that the position of the space target in the high-sensitivity image remains unchanged.

[0060] The high-sensitivity camera of the optical measurement equipment continuously shoots the space target to obtain n frames of continuous high-sensitivity images.

[0061] S2: Count the grayscale values ​​in the row and column directions in the high-sensitivity image sequence to obtain the brightness distribution.

[0062] Assume that the width and height of each frame of high-sensitivity image are W and H, the grayscale value of the pixel point (x, y) is Img_i(x, y), x is the horizontal coordinate of the pixel point in the high-sensitivity image, and y is the vertical coordinate of the pixel point in the high-sensitivity image.

[0063] Firstly, the grayscale values ​​in the row direction and the column direction of each continuous frame of high-sensitivity image are counted, and the grayscale values ​​in the row direction and the column direction are counted row by row and column by column.

[0064] Take the i-th frame of high-sensitivity image as an example:

[0065] The grayscale values ​​of the y-th line of the i-th frame high-sensitivity image are summed to obtain the line grayscale value Img_i_line(y).

[0066] Img_i_line(y)=Img_i(0,y)+Img_i(1,y)+…+Img_i(W-1,y); wherein Img_i(0,y), Img_i(1,y), …, Img_i(W-1,y) respectively represent the grayscale values ​​of pixel (0,y), pixel (1,y), …, pixel (W-1,y) in the i-th frame of high-sensitivity image.

[0067] The gray value statistics of other rows of the i-th frame high-sensitivity image can be obtained similarly.

[0068] The grayscale values ​​of the x-th column of the i-th frame high-sensitivity image are summed to obtain the column grayscale value Img_i_column(x).

[0069] Img_i_column(x)=Img_i(x,0)+Img_i(x,1)+…+Img_i(x,H-1); where Img_i(x,0), Img_i(x,1), …, Img_i(x,H-1) represent the grayscale values ​​of pixel (x,0), pixel (x,1), …, pixel (x,H-1) in the i-th frame of high-sensitivity image, respectively.

[0070] The gray value statistics of other columns of the i-th frame high-sensitivity image can be obtained similarly.

[0071] Secondly, the row grayscale values ​​of each row of all high-sensitivity images are summed to obtain the total row grayscale value, and the row grayscale values ​​of each column of all high-sensitivity images are summed to obtain the total column grayscale value.

[0072] Take the pixel (x, y) as an example:

[0073] The row grayscale values ​​of the yth row of the n-frame high-sensitivity image are summed to obtain the row grayscale total value Img_line(y).

[0074] Img_line(y)=Img_1_line(y)+Img_2_line(y)+…+Img_n_line(y); wherein, Img_1_line(y), Img_2_line(y), …, Img_n_line(y) represent the grayscale value of the yth line of the high-sensitivity images of the 1st to nth frames, respectively.

[0075] The same can be achieved by summing up the grayscale values ​​of other rows of the n-frame high-sensitivity image.

[0076] The column grayscale values ​​of the x-th column of the n-frame high-sensitivity image are summed to obtain the column grayscale total value Img_column(x).

[0077] Img_column(x)=Img_1_column(x)+Img_2_column(x)+…+Img_n_column(x); wherein Img_1_column(x), Img_2_column(x), …, Img_n_column(x) represent the column grayscale values ​​of the xth column of the high-sensitivity images of the 1st frame to the nth frame, respectively.

[0078] The same can be done for summing up the grayscale values ​​of other columns of n frames of high-sensitivity images.

[0079] Through the above-mentioned statistics of the grayscale values ​​of rows and columns, the brightness distribution of the high-sensitivity image is obtained, which is convenient for the subsequent distinction between spatial targets and noise.

[0080] Since the space target is a bright target, the grayscale value of its image is relatively high. Assuming the coordinates of the space target in the high-sensitivity image are (xn, yn), it will inevitably lead to the sum of the grayscale values ​​of Img_i_line(xn), Img_i_column(yn) and its adjacent row and column positions being significantly higher than the sum of the grayscale values ​​of other positions.

[0081] It should be noted that (xn,yn) represents a plurality of pixel points occupied by the space target in the high-sensitivity image, n is an independent variable, xn represents the horizontal coordinate of the pixel points occupied by the space target, and yn represents the vertical coordinate of the pixel points occupied by the space target.

[0082] S3: Determine the coordinates of the space target in the high-sensitivity image sequence based on the brightness distribution.

[0083] According to the gray value, count the top 0.5% of the pixels in the total gray value of the column Img_column(x) and the total gray value of the row Img_line(y), and observe the values ​​of x and y. Remove the discontinuous pixels in the statistical results to obtain the horizontal coordinate set {x1, x2, ... xa} and the vertical coordinate set {y1, y2, ... yb}, then:

[0084] xn=(x1+x2+…+xa) / a;

[0085] yn=(y1+y2+…+yb) / b.

[0086] The purpose of the above operation is to select the rows and columns where the brightest pixels in the high-sensitivity image are located, so the average grayscale value of the pixel rows and columns is used for measurement. The specific operation is to select the brightest top 0.5% of the pixels in the total column grayscale value Img_column(x) and the total line grayscale value Img_line(y) for grayscale value statistics. Since there are some bad pixels like snowflakes in infrared images, these noises often only occupy one or two pixels, but the grayscale values ​​are often relatively high (similar to white points), and non-continuous x and y values ​​represent these noises. Therefore, removing discontinuous pixels in the statistical results can eliminate the interference of these noises on the grayscale value statistics.

[0087] The mean of all grayscale values ​​of a row or a column in n consecutive high-sensitivity images is calculated to measure whether there is a space target in this row or column. Since the grayscale value of the space target is brighter, if there is a space target, the mean grayscale value of this row or column will be higher, so as to obtain the row coordinates and vertical coordinates of the space target, and then quickly locate the position of the space target in the high-sensitivity image.

[0088] S4: Adjust the exposure of the high-sensitivity camera in the optical measurement device, re-image the space target, obtain a 16-bit original image, and make the brightness of the pixels around the space target in the 16-bit original image reach the brightness threshold.

[0089] By adjusting the exposure of the high-sensitivity camera and the brightness of the high-sensitivity image, the grayscale value of the pixel points around the space target (xn, yn) is close to the saturation state. The present invention uses a 16-bit high-sensitivity image, the saturation grayscale value is 65535, and the brightness threshold is 50000. The exposure of the high-sensitivity camera is controlled so that the brightness of the four pixel areas around the coordinate point (xn, yn) of the space target is above 50000. At this time, the image of the space target in the high-sensitivity image is brighter, and the image of the noise is also enhanced, but the brightness of the noise part is relatively low compared with the brightness of the space target.

[0090] S5: The high 8-bit data of the 16-bit original image is selected and the grayscale is stretched at equal intervals to obtain two 8-bit images.

[0091] (1) Select the high 8-bit data of the 16-bit original image

[0092] Select the high 8-bit data of the 16-bit original image to obtain an 8-bit single-channel image. Due to the brightness difference between the space target and the noise, the space target can still be clearly imaged in the high 8-bit data, but the noise information will be completely filtered out due to its low brightness. But at the same time, the images of some faint celestial bodies may also be suppressed. Suppose the 16-bit original image is Img0, and the grayscale value at the pixel point (x, y) in the original image Img0 is recorded as Img0(x, y). After taking the high 8-bit data, the obtained 8-bit single-channel image is Img1, and the grayscale value at the pixel point (x, y) in the single-channel image Img1 becomes Img1(x, y):

[0093]

[0094] (2) Stretching the grayscale of the 16-bit original image with equal spacing

[0095] First, the grayscale values ​​of all pixels in the original image Img0 are counted to obtain a set of pixels in the grayscale value range of 0-65535.

[0096] Then, count the number of valid grayscale values ​​in the pixel set, record it as count, and record these valid grayscale values ​​in ascending order. The definition of valid grayscale value is: traverse the image point by point, count a grayscale value val in the range of 0-65535, and determine whether there is a pixel with grayscale value val in the high-sensitivity image. If so, val is called a valid grayscale value, otherwise, val is called an invalid grayscale value.

[0097] At this time, the pixels in the original image Img0 are transformed to generate an 8-bit single-channel image Img2: the grayscale value at the pixel point (x, y) in the single-channel image Img2 becomes Img2(x, y). Suppose in the original image Img0, if Img0(x, y) ranks t in the ascending order among all valid grayscale values, then:

[0098]

[0099] S6: Fuse the two 8-bit images to obtain a denoised high-sensitivity image.

[0100] Since both the single-channel image Img1 and the single-channel image Img2 are 8-bit images, the two are fused to obtain a denoised high-sensitivity image Img3.

[0101] The grayscale value of the high-sensitivity image Img3 is Img3(x,y), and the calculation formula is:

[0102]

[0103] Among them, μ is the fusion parameter, which is controlled by manual setting.

[0104] Since the method of the present invention does not involve complex matrix operations and does not involve data transmission between GPUs, the method itself is less time-consuming, consuming only milliseconds, ensuring a fast noise suppression speed and meeting the high real-time requirements of optical measurement equipment.

[0105] After step S6, the method further includes the following steps:

[0106] S7: Use a stretching method to stretch the high-sensitivity image Img3.

[0107] The noise in the denoised high-sensitivity image Img3 has been weakened to a certain extent. By stretching the high-sensitivity image Img3, the grayscale difference between the spatial target and the noise can be further enlarged.

[0108] The stretching method may be a linear stretching method or a nonlinear stretching method.

[0109] Taking the nonlinear stretching method as an example, the nonlinear processing function used by the nonlinear stretching method is as follows Figure 2 As shown, the transformation function curve selected in the present invention is represented by y=f(x), and y=f(x) is an "S"-shaped curve. Figure 2 From the analysis, we can see that when the gray value x in the image before the transformation is high, the result f(x) after the transformation is greater than x, and when the gray value x in the image before the transformation is low, the result f(x) after the transformation is less than x, that is, after this transformation, the gray value of the bright part of the space target in the image increases, and the corresponding white area will be whiter, and the gray value of the background and noise area in the image decreases, and the corresponding black will be darker. This further suppresses the noise.

[0110] The final image Img_dst after processing is:

[0111]

[0112] At this point, the final denoised image Img_dst can be obtained. Figure 3 This is the image effect before denoising. Figure 4 This is the image effect after denoising.

[0113] After image fusion, the noise will be suppressed, the noise intensity will be weakened, and it will be closer to the background grayscale. Image stretching can make the difference between light and dark more obvious, so the noise will be further suppressed.

[0114] The present invention utilizes a method in which there is a certain brightness difference between the space target and the image background. First, the exposure is adjusted by a high-sensitivity camera to increase the brightness of the high-sensitivity image. At this time, the brightness difference between the space target and the image background will become larger. Then, by selecting the high eight-bit data, the noise in the background is removed to complete the initial suppression of the noise. After that, the two 8-bit images are weighted fused, and the background information of the fused image is restored. Finally, the image is stretched to further enhance the brighter space target and weaken the background noise. Because the dim targets such as dim stars have a certain brightness difference with the background, they will be retained on the background through image fusion and stretching and will not be completely removed, while the background noise will be suppressed.

[0115] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps described in the disclosure of the present invention can be performed in parallel, sequentially or in different orders, as long as the desired results of the technical solution disclosed in the present invention can be achieved, and this document does not limit this.

[0116] The above specific implementations do not constitute a limitation on the protection scope of the present invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for suppressing background noise of a high-sensitivity image of a space target, characterized in that: The steps include: S1: Use optical measurement equipment to track and image space targets to obtain continuous high-sensitivity image sequences; S2: Count the grayscale values ​​in the row and column directions in the high-sensitivity image sequence to obtain the brightness distribution; S3: Determine the coordinates of the space target in the high-sensitivity image sequence based on the brightness distribution; S4: adjusting the exposure of the high-sensitivity camera in the optical measurement device, re-imaging the space target, obtaining a 16-bit original image, and making the brightness of the pixel points around the space target in the 16-bit original image reach the brightness threshold; S5: simultaneously select the high 8-bit data of the 16-bit original image and stretch the grayscale with equal spacing to obtain two 8-bit images; S6: Fuse the two 8-bit images to obtain a denoised high-sensitivity image.

2. The method for suppressing background noise of a space target high-sensitivity image according to claim 1, characterized in that: Assume that the high-sensitivity image sequence includes n frames of high-sensitivity images, the width and height of each frame of high-sensitivity image are W and H respectively, and (x, y) is the horizontal coordinate and vertical coordinate of the pixel point in the high-sensitivity image; In step S2, the process of counting the grayscale values ​​in the row direction and the column direction in the high-sensitivity image sequence is as follows: Sum the grayscale values ​​of each row of the i-th frame high-sensitivity image to obtain the row grayscale value Img_i_line(y), Img_i_line(y)=Img_i(0,y)+Img_i(1,y)+…+Img_i(W-1,y), Img_i(0,y), Img_i(1,y), …, Img_i(W-1,y) respectively represent the grayscale values ​​of pixel (0,y), pixel (1,y), …, pixel (W-1,y) in the i-th frame high-sensitivity image; Sum the grayscale values ​​of each column of the i-th frame of high-sensitivity image to obtain the column grayscale value Img_i_column(x), Img_i_column(x)=Img_i(x,0)+Img_i(x,1)+…+Img_i(x,H-1), where Img_i(x,0), Img_i(x,1), …, Img_i(x,H-1) represent the grayscale values ​​of pixel (x,0), pixel (x,1), …, pixel (x,H-1) in the i-th frame of high-sensitivity image, respectively; Sum the grayscale value of each line of the n-frame high-sensitivity image to obtain the total grayscale value Img_line(y), Img_line(y)=Img_1_line(y)+Img_2_line(y)+…+Img_n_line(y), Img_1_line(y), Img_2_line(y), …, Img_n_line(y) represent the grayscale value of each line of the 1st to the nth frame high-sensitivity images respectively; The column grayscale values ​​of each column of n frames of high-sensitivity images are summed to obtain the total column grayscale value Img_column(x), Img_column(x)=Img_1_column(x)+Img_2_column(x)+…+Img_n_column(x), Img_1_column(x), Img_2_column(x), …, Img_n_column(x) represent the column grayscale values ​​of each column of the 1st to nth frames of high-sensitivity images, respectively.

3. The method for suppressing background noise of a space target high-sensitivity image according to claim 2, characterized in that: Assume that the coordinates of the space target in the high-sensitivity image are (xn,yn); In step S3, the process of determining the coordinates of the space target in the high-sensitivity image sequence based on the brightness distribution is as follows: According to the gray value, the top 0.5% of the pixels in the total gray value of the column Img_column(x) and the total gray value of the row Img_line(y) are counted, and the discontinuous pixels in the statistical results are removed to obtain the abscissa set {x1,x2,…xa} and the ordinate set {y1,y2,…yb}, then: xn=(x1+x2+…+xa) / a; yn=(y1+y2+…+yb) / b.

4. The method for suppressing background noise of a space target high-sensitivity image according to any one of claims 1 to 3, characterized in that: In step S4, the brightness threshold is set to 50,000.

5. The method for suppressing background noise of a space target high-sensitivity image according to claim 4, characterized in that: Assume that the 16-bit original image is Img0, and the grayscale value of the pixel (x, y) in the original image Img0 is Img0(x, y). Then in step S5, the high 8-bit data of the original image Img0 is selected, and the obtained 8-bit image is Img1. The grayscale value Img0(x, y) of the pixel (x, y) in the 8-bit image Img1 is converted to Img1(x, y) by the following formula:

6. The method for suppressing background noise of a space target high-sensitivity image according to claim 5, characterized in that: In step S5, the process of stretching the original image Img0 with equal spacing in grayscale is as follows: The grayscale values ​​of all pixels in the original image Img0 are counted to obtain a set of pixels in the grayscale value range of 0-65535; Count the number of valid pixels in the pixel set, record it as count, and record it in ascending order; The original image Img0 is transformed to obtain an 8-bit image Img2. The grayscale value Img0(x,y) of the pixel (x,y) in the 8-bit image Img2 is converted to Img2(x,y) by the following formula: Among them, t is the ranking of the gray value Img0(x,y) of the pixel (x,y) in the original image Img0 in the valid gray values ​​in ascending order.

7. The method for suppressing background noise of a space target high-sensitivity image according to claim 6, characterized in that: In step S6, the two 8-bit images are fused to obtain a denoised high-sensitivity image Img3. The grayscale value of the high-sensitivity image Img3 is Img3(x, y), and the calculation formula is: Among them, μ is the fusion parameter.

8. The method for suppressing background noise of a space target high-sensitivity image according to claim 7, characterized in that: After step S6, the method further includes the following steps: S7: Use a stretching method to stretch the high-sensitivity image Img3.

9. The method for suppressing background noise of a space target high-sensitivity image according to claim 8, characterized in that: The stretching method is a linear stretching method or a nonlinear stretching method.

Images