Star point image processing method in complex background

Abstract

The star sensor acquires a star point image through a detector, a data processor scans the obtained star point image using a 3*3 template; judges whether the gray value of the pixel point in the center of the template is greater than an image threshold T1 and whether the gray values of the 8 pixels around the pixel point are all less than the image threshold T1, and obtains a star point image after removing isolated points; the data processor divides the star point image after removing isolated points into blocks, divides the size of a sub-block into 32*32 pixels, and then calculates the image mean value in each sub-block; judges the mean value of each sub-block, and filters out the background and noise with high gray value; performs a pixel traversal operation on the star point image after removing isolated noise using the determined image threshold T1, and sends the pixels not less than the threshold T1 to a target extraction module. The star point image processing method under a complex background can judge and filter the star point image under a complex background, and can stably output the attitude information of the star sensor.

Description

Technical Field

[0001] This invention relates to the field of astronomical navigation technology, specifically to a method for processing star point images against a complex background. Background Technology

[0002] A star sensor is a high-precision attitude-sensing measurement instrument that determines a carrier's attitude by detecting stars at different positions in the celestial sphere. It is currently the most widely used attitude measurement device in the aerospace field. With the successful launch of numerous artificial satellites in space, these satellites interfere with the star sensor's imaging system by reflecting sunlight, making the space environment increasingly complex. Furthermore, it is subject to interference from lunar and atmospheric stray light. In harsh environments, the background of star images produced by star sensors is extremely complex, manifested in uneven grayscale distribution, stars being obscured by background noise, and the presence of isolated points of noise such as thermal noise. These characteristics severely impact star target extraction. Traditional star target extraction methods often fail to extract a large number of stars, resulting in inconsistent and accurate output of the centroid of the target stars. This hinders proper star image recognition and attitude calculation, ultimately causing the star sensor to become "lost" in space and lose its normal navigation function. Summary of the Invention

[0003] This invention addresses the problem that traditional star image processing methods often fail to extract star targets in complex backgrounds, by providing a star image processing method applicable to star sensors in complex backgrounds.

[0004] A method for processing star-shaped images against a complex background, characterized in that the method includes the following steps:

[0005] Step 1: The star sensor acquires star images through the detector, and the data processor scans the obtained star images using a 3×3 template;

[0006] Determine whether the gray value of the center pixel of the template is greater than the image threshold T1, and whether the gray values ​​of the 8 surrounding pixels are all less than the image threshold T1, to obtain the star image after removing isolated points;

[0007] Step 2: The data processor divides the star image obtained in Step 1 after removing isolated points into blocks. The size of each sub-block can be 32×32 pixels. Then, the average value of the image within each sub-block is calculated.

[0008] Step 3: Determine whether the mean value of each sub-block is greater than the threshold T2, and further filter out the background and noise of the high-brightness grayscale values;

[0009] Step 4: Using the image threshold T1 determined in Step 2 and Step 3, perform a pixel traversal operation on the star image after removing isolated point noise, and send the pixels not less than the threshold T1 to the target extraction module.

[0010] Preferably, in step one, it is determined whether the gray value of the center pixel of the template is greater than the image threshold T1, and whether the gray values ​​of the surrounding 8 pixels are all less than the image threshold T1.

[0011] If so, this center pixel is isolated noise, and the gray value of the right pixel or any of the surrounding pixels is used to replace this center pixel.

[0012] If not, then the center pixel is a non-noise point, and the grayscale value of this pixel remains unchanged.

[0013] Preferably, in step two, the values ​​of the row counter and column counter are determined, the grayscale sum is calculated for every 32 rows and 32 columns, and then the grayscale sum is shifted right by 10 bits at the falling edge of the field, which is the mean of each sub-block.

[0014] The image threshold T1 within a sub-block is calculated using the following formula:

[0015] T1=v+aσ

[0016] In the formula, v is the average gray value of the star image sub-block, a is the scale factor, and σ is the standard deviation of the star image sub-block.

[0017] Preferably, in step three, the mean value v of each sub-block is judged. When the mean value of the sub-block is greater than the threshold T2, it is considered that the sub-block is severely polluted by noise and the star points are completely buried in the sub-block.

[0018] Preferably, the threshold T2 is set according to the detection capability of the star sensor, specifically the average gray value of a 0 magnitude star imaged in the star sensor imaging unit plus the standard deviation of the entire frame image.

[0019] The beneficial effects of this invention are as follows: A method for processing star images under complex backgrounds can identify and filter noise in star images under complex backgrounds, and is applicable to star images under ordinary backgrounds. Furthermore, it can still stably output star sensor attitude information when complex background conditions occur in the star sensor's field of view. This invention is easy to implement on hardware platforms, and the method is simple, reliable, and effective, which can greatly improve the applicability of star sensors. Attached Figure Description

[0020] Figure 1 A schematic diagram illustrating the principle of star image processing methods against complex backgrounds;

[0021] Figure 2 This is a flowchart of the process for removing isolated point noise in star point image processing methods under complex backgrounds.

[0022] Figure 3 A schematic diagram illustrating the principle of eight-connectivity analysis of star targets in star image processing methods under complex backgrounds; Detailed Implementation

[0023] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0024] The specific implementation of the present invention will be described in detail below with reference to specific embodiments:

[0025] Step 1: The star image acquired by the star sensor usually contains isolated point noise. The data processor uses a 3×3 template to scan the entire frame image and determines whether the gray value of the center pixel of the template is greater than the image threshold T1, and whether the gray values ​​of the surrounding 8 pixels are all less than the image threshold T1.

[0026] If so, then this center pixel is isolated noise, and its grayscale value is replaced by the pixel to its right or any of the surrounding pixels.

[0027] If not, then the center pixel is a non-noise point, and its grayscale value remains unchanged. By traversing the entire image frame in this way, the goal of removing isolated point noise is ultimately achieved.

[0028] The important role of this step is that a large number of isolated point noises will occupy the computing and storage resources of the data processor. Firstly, removing isolated points from the star image will greatly reduce the amount of computation in data processing, prevent data blockage, and improve the computing efficiency of the data processor.

[0029] Step 2: The data processor divides the star image obtained in Step 1 after removing isolated points into blocks. Taking the whole frame image size as 2048×2048 as an example, the star points in the star sensor detector are usually within 15×15 pixels in size. Therefore, the sub-block size can be divided into 32×32 pixels. Then, the image mean v in each sub-block is calculated.

[0030] The specific method is as follows: determine the values ​​of the row counter and column counter, calculate the grayscale sum for every 32 rows and 32 columns, and then right-shift the grayscale sum by 10 bits at the falling edge of the field. This gives the mean value for each sub-block.

[0031] Finally, the image threshold T1 within the sub-block is calculated using the following formula:

[0032] T1=v+aσ

[0033] In the formula, v is the average gray value of the star map sub-block, a is the scale factor, and σ is the standard deviation of the star map sub-block. Based on the properties of the normal distribution and referring to experimental results, a is taken as 3.

[0034] The function of the threshold T1 is:

[0035] 1. Used for threshold segmentation and connectivity analysis of each pixel within each sub-block;

[0036] 2. The operation of judging the gray value of the center pixel of the template during isolated point removal in step one is used to distinguish whether it is isolated point noise.

[0037] Step 3: In order to further filter out background and noise with high gray values, such as large light spots like the sun, moon, and ground atmosphere, these noises are characterized by very high average gray values ​​within sub-blocks, usually close to or already saturated.

[0038] Therefore, the mean value v of each sub-block is judged. When the mean value of a sub-block is greater than the threshold T2, it is considered that the sub-block is severely contaminated by noise, and the star points are completely buried in the sub-block. Then, the image threshold T1 of this sub-block is set to 4095 (taking a 12-bit image as an example). In this way, the target extraction function module will automatically filter out this sub-block, thereby achieving the purpose of filtering out large-area spot noise.

[0039] The threshold T2 is set according to the detection capability of the star sensor, specifically the average gray value of the image of a 0 magnitude star in the star sensor imaging unit plus the standard deviation of the entire frame image.

[0040] Step 4: Using the image threshold T1 determined in Steps 2 and 3, perform a pixel traversal operation on the star map after removing isolated point noise. Pixels not smaller than the threshold T1 are sent to the target extraction module. The target extraction module uses an 8-connectivity analysis method, the principle of which is as follows: Figure 3 As shown, the specific formula for calculating the centroid is as follows:

[0041]

[0042] The obtained star targets are further filtered. Since the star sensor optical system uses defocus imaging, the size of the point targets formed by stars on the star sensor imaging device is usually greater than 2×2 and less than 15×15. Therefore, in this embodiment, isolated points, random noise and other false targets with less than 4 pixels are deleted, as well as large spot targets with more than 225 pixels are deleted. After this step, the star sensor completes the processing of star point images under complex backgrounds and achieves continuous and accurate output of the centroid results of star point targets.

[0043] The beneficial effects of this invention are as follows: A method for processing star images under complex backgrounds can identify and filter noise in star images under complex backgrounds, and is applicable to star images under ordinary backgrounds. Furthermore, it can still stably output star sensor attitude information when complex background conditions occur in the star sensor's field of view. This invention is easy to implement on hardware platforms, and the method is simple, reliable, and effective, which can greatly improve the applicability of star sensors.

[0044] Optionally, in an exemplary embodiment, a computer-readable storage medium is shown storing a computer program, characterized in that the computer program, when executed, implements the method described in Embodiment 1.

[0045] 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, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for processing star point images against a complex background, characterized in that, The method includes the following steps: Step 1: The star sensor acquires star images through the detector, and the data processor scans the obtained star images using a 3×3 template; Determine whether the gray value of the center pixel of the template is greater than the image threshold T1, and whether the gray values ​​of the 8 surrounding pixels are all less than the image threshold T1, to obtain the star image after removing isolated points; Step 2: The data processor divides the star image obtained in Step 1 after removing isolated points into blocks, with each sub-block being 32×32 pixels in size, and then calculates the image mean within each sub-block; Step 3: Determine whether the mean value of each sub-block is greater than the threshold T2, and further filter out the background and noise with high-brightness grayscale values; Step 4: Using the image threshold T1 determined in Step 2, perform a pixel traversal operation on the star image after removing isolated point noise, and send the pixels not less than the threshold T1 to the target extraction module. In step two, the values ​​of the row counter and column counter are determined, and the grayscale sum is accumulated for every 32 rows and 32 columns. Then, at the falling edge of the field, the grayscale sum is shifted 10 bits to the right, which is the mean of each sub-block. The image threshold T1 within a sub-block is calculated using the following formula: T1=v+aσ In the formula, v is the average gray value of the star image sub-block, a is the scale factor, and σ is the standard deviation of the star image sub-block; In step three, the mean value of each sub-block is judged. When the mean value of a sub-block is greater than the threshold T2, it is considered that the sub-block is severely polluted by noise and the star points are completely buried in the sub-block. The threshold T2 is set according to the detection capability of the star sensor, specifically the average gray value of a 0 magnitude star in the star sensor imaging unit plus the standard deviation of the entire frame image.

2. The star point image processing method under complex backgrounds as described in claim 1, characterized in that, In step one, it is determined whether the gray value of the center pixel of the template is greater than the image threshold T1, and whether the gray values ​​of the surrounding 8 pixels are all less than the image threshold T1. If so, this center pixel is isolated noise, and the gray value of the right pixel or any of the surrounding pixels is used to replace this center pixel. If not, then the center pixel is a non-noise point, and the grayscale value of this pixel remains unchanged.

Images