A method, system, and medium for detecting impact craters on multi-scale star catalogs.

Abstract

This invention designs a multi-scale star surface impact crater detection method, system, and medium. The method includes the following steps: S1: Acquire star surface images and segment them into image patches of different scales; S2: Construct multiple impact crater detection sub-models; S3: Train each sub-model using star surface image patches of different scales; S4: Input the segmented star surface image to be detected into each pre-trained sub-model to obtain the detection results; S5: Select the bounding boxes with high confidence and large area among all candidate bounding boxes of each impact crater as the final detection result for that impact crater. This invention can simultaneously identify impact craters of different scales in star surface images with high accuracy, especially improving the accuracy and efficiency of identifying small impact craters, providing more accurate and comprehensive data support for lunar science research, and promoting the in-depth development of lunar exploration and geological research.

Description

Technical Field

[0001] This invention belongs to the field of remote sensing image processing technology, and particularly relates to a method, system and medium for detecting multi-scale star surface impact craters. Background Technology

[0002] In the fields of planetary exploration and geological research, the identification and extraction of impact craters on planetary surfaces has always been an important research direction. Impact craters are not only one of the most prominent geomorphic features on planetary surfaces, but they also carry rich geological information, which is of great significance for understanding the evolutionary history, geological structure, and surface processes of planets.

[0003] Traditional impact crater identification methods primarily rely on manual identification and visual interpretation. Researchers manually mark and record the location, size, and other relevant parameters of impact craters by directly observing and analyzing images or photographs of the planetary surface. While this method can meet research needs to some extent, it is extremely time-consuming and labor-intensive, and the identification results are often subjective and inconsistent, making it difficult to guarantee data quality and accuracy.

[0004] To overcome this limitation, existing technologies typically utilize computer vision and image processing techniques, such as image analysis, data fusion, machine learning, and training deep learning network models, to process and analyze planetary surface images, thereby achieving automatic identification and extraction of impact craters. While these methods have achieved some success and improved the efficiency and accuracy of impact crater identification to a certain extent, the following problems still exist:

[0005] 1. Poor performance of multi-scale identification: Impact craters on the surface of planets vary in size, ranging from a few meters to hundreds of kilometers in diameter. Traditional single-scale identification methods are difficult to adapt to impact craters of different sizes at the same time, especially when identifying small impact craters.

[0006] 2. Imbalanced sample problem: There are relatively few large impact craters on the surface of planets, while there are many small impact craters. This imbalanced sample problem causes machine learning models to tend to identify large, distinctive impact craters during training, while ignoring or misjudging small impact craters.

[0007] 3. Insufficient recognition accuracy: Traditional recognition methods, especially those based on simple image processing techniques, often struggle to achieve high accuracy when dealing with small impact craters with indistinct features or irregular shapes. This can lead to a large number of small impact craters being missed or misidentified.

[0008] 4. Inefficiency: Methods relying on manual identification or simple automation are inefficient when processing large-scale planetary surface data, and cannot meet the requirements of modern planetary exploration missions for data processing speed and accuracy.

[0009] Therefore, there is an urgent need to develop a detection method that can simultaneously identify impact craters of multiple scales on the surface of a planet, in order to solve the problem that existing detection methods cannot simultaneously identify impact craters of different sizes on the surface of a planet, especially when identifying small impact craters, due to insufficient accuracy and robustness. Summary of the Invention

[0010] The purpose of this invention is to solve the above-mentioned technical problems and provide a multi-scale impact crater detection method, system and medium. By adopting multi-scale detection algorithms, data augmentation and image processing technology, it overcomes the problems of insufficient recognition accuracy, sample imbalance and poor multi-scale recognition effect faced by existing technologies in identifying impact craters. It improves the accuracy and recognition efficiency of feature extraction of small impact craters on planetary surfaces, provides more accurate and comprehensive data support for lunar science research, and promotes the in-depth development of lunar exploration and geological research.

[0011] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0012] A method for detecting impact craters on a multi-scale star catalog includes the following steps:

[0013] S1: Acquire star surface image data containing impact craters, and segment the acquired star surface images at different scales to obtain star surface image patches of different scales; star surface image patches of different scales contain impact crater targets of different diameter ranges respectively.

[0014] S2: Construct a multi-scale impact crater detection model, which includes multiple impact crater detection sub-models based on YOLOv5;

[0015] S3: Multiple impact crater detection sub-models are trained using star catalog image patches of different scales to obtain multiple sub-models for detecting and recognizing impact crater targets with different diameter ranges.

[0016] S4: Obtain the star catalog image to be detected, divide it into image blocks of different scales, and input them into the sub-models of the corresponding scale ranges to obtain all candidate bounding boxes of impact crater targets in the star catalog image to be detected.

[0017] S5: Sort all candidate bounding boxes for each impact crater target based on the area of ​​the candidate bounding boxes, calculate the IoU value between the candidate bounding box with the largest area and other candidate bounding boxes, and select the candidate bounding box with the largest area as the final detection result of the impact crater target among all candidate bounding boxes with IoU values ​​exceeding a predetermined IoU threshold.

[0018] In some embodiments of the present invention, S5 specifically includes the following steps:

[0019] S51: Define the predetermined IoU threshold τ and the candidate bounding box list B for each impact crater target:

[0020] B = {B1, B2, ..., B} N};

[0021] Each candidate bounding box contains location information;

[0022] S52: Sort the candidate bounding boxes in list B based on their area;

[0023] S53: Select the candidate bounding box B with the largest current area from list B. i , i∈[1,N];

[0024] Calculate the other candidate bounding boxes in list B and B i The IoU value, which will include all values ​​related to B i Candidate bounding boxes whose IoU values ​​exceed a predetermined IoU threshold τ are all added to a predetermined empty set G;

[0025] If set G is not empty after the calculation, then in B i Select the candidate bounding box B{max} with the largest area from set G and add it to a predetermined result list D;

[0026] S54: Remove the candidate bounding box with the largest current area from list B and all candidate bounding boxes from set G;

[0027] S55: Repeat steps S53-S54 until all candidate bounding boxes in list B are removed;

[0028] S56: Output the candidate bounding boxes in list D as the final detection result for the impact crater target.

[0029] In some embodiments of the present invention, step S1, which involves segmenting each star catalog image at different scales, specifically includes the following steps:

[0030] Set up three different sized sliding windows;

[0031] Three different sized sliding windows start from the top left corner of each star catalog image and slide sequentially from left to right and from top to bottom, cropping out star catalog image blocks of the corresponding window size from the image during the sliding process.

[0032] In some embodiments of the present invention, three different sizes of sliding windows are 320×320 pixels, 640×640 pixels and 1280×1280 pixels respectively.

[0033] In some embodiments of the present invention, the following steps are also included:

[0034] Set a first diameter threshold and a second diameter threshold; the second diameter threshold is greater than the first diameter threshold.

[0035] The image patches cropped by the three different sizes of sliding windows respectively contain impact crater targets with diameters below the first diameter threshold, impact crater targets between the first diameter threshold and the second diameter threshold, and impact crater targets above the second diameter threshold.

[0036] In some embodiments of the present invention, the first diameter threshold is 50 pixels and the second diameter threshold is 100 pixels.

[0037] In some embodiments of the present invention, the following steps are also included:

[0038] Set a predetermined scale threshold;

[0039] In step S1, after obtaining star catalog image patches of different scales, for image patches with side lengths below a predetermined scale threshold, they are divided into multiple sub-regions of the same size, each sub-region is magnified, and the magnified image patch is further segmented based on the division of each sub-region.

[0040] In some embodiments of the present invention, the value range of the predetermined scale threshold is 320 pixels to 640 pixels.

[0041] Some embodiments of the present invention further provide a multi-scale star catalog impact crater detection system, characterized in that it includes:

[0042] At least one processor;

[0043] At least one memory for storing at least one program;

[0044] When at least one program is executed by at least one processor, the at least one processor implements a multi-scale star catalog impact crater detection method.

[0045] Some embodiments of the present invention further provide a storage medium storing a processor-executable program, characterized in that: the processor-executable program, when executed by a processor, is used to implement a multi-scale star catalog impact crater detection method.

[0046] The beneficial effects of this invention are as follows:

[0047] 1. This invention constructs multiple impact crater detection sub-models and trains them using image patches of different scales to obtain multiple sub-models for detecting and recognizing impact crater targets with different diameter ranges. This enables the trained multi-scale impact crater detection model to detect impact craters with different diameter ranges simultaneously, thereby improving the accuracy of the model in recognizing impact craters of various sizes.

[0048] 2. In this invention, images containing small-scale impact crater targets are sliced, magnified, and then sliced ​​again, which improves the clarity of small impact crater features in the image and thus improves the model's ability to extract features from small impact craters.

[0049] 3. By evaluating the overlap and confidence between different candidate bounding boxes, this invention prioritizes the retention of prediction results with high confidence and large area, which can obtain more accurate and complete impact crater identification results. While maintaining detection accuracy, it reduces redundant predictions and improves identification efficiency.

[0050] 4. This invention overcomes the problems of insufficient recognition accuracy, sample imbalance, and poor multi-scale recognition effect faced by existing technologies in identifying impact craters of different scales on the planetary surface by constructing a multi-scale impact crater detection model. In particular, it improves the accuracy and efficiency of feature extraction and recognition of small impact craters on the planetary surface, providing more accurate and comprehensive data support for lunar science research and promoting the in-depth development of lunar exploration and geological research. Attached Figure Description

[0051] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0052] Figure 1 A flowchart of a multi-scale star catalog impact crater detection method;

[0053] Figure 2 This is a schematic diagram of the star catalog image data processing process in this embodiment;

[0054] Figure 3 This is a schematic diagram of the sliding window clipping process;

[0055] Figure 4 This is a schematic diagram of the process of enlarging and segmenting a small-scale image.

[0056] Figure 5 This is a schematic diagram of the redundant impact crater detection results of the multi-scale impact crater detection model.

[0057] Among them, (a), (b), (c), and (d) are the redundancy detection results of the same impact crater in different slices, respectively;

[0058] Figure 6 This is a flowchart of a method for handling redundant impact craters in the prior art;

[0059] Figure 7 This is a flowchart of the redundant impact crater treatment method provided in this embodiment;

[0060] Figure 8 A statistical chart showing the results of the detection method provided in this embodiment identifying 500,000 impact craters in the Chang'e-4 landing area;

[0061] Figure 9 This is a schematic diagram of the detection results for identifying multi-scale impact craters using the detection method provided in this embodiment. Detailed Implementation

[0062] To make the objectives, technical solutions, and advantages of this application clearer, the application is described and illustrated 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 scope of this application. All other embodiments obtained by those skilled in the art based on the embodiments provided in this application without inventive effort are within the scope of protection of this application.

[0063] It should be noted that the terminology used herein is for the purpose of describing particular implementations only and is not intended to limit the exemplary implementations according to this application. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. Furthermore, it should be understood that the terms “comprising” and “having”, and any variations thereof, are intended to cover non-exclusive inclusion, for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such process, method, product, or apparatus.

[0064] Where there is no conflict, the embodiments and features in the embodiments of the present invention can be combined with each other.

[0065] The technical solution of the present invention will be described in detail below with reference to specific embodiments and accompanying drawings.

[0066] As attached Figure 1 - Appendix Figure 9 As shown in the illustrative embodiment of the multi-scale star surface impact crater detection method, system and medium of the present invention, the method is used to detect and identify impact craters of different scales on the moon, and the method includes the following steps.

[0067] S1: Data acquisition and preprocessing.

[0068] First, high-resolution image data of the lunar surface collected by lunar probes such as Chang'e-4 were collected and compiled into a dataset. Based on the image data in the dataset, a detailed study and scale classification of impact craters on the lunar surface were conducted, with particular attention paid to the characteristics and distribution patterns of small impact craters in the images.

[0069] Multiple high-resolution star surface images containing impact craters were obtained from this dataset. A sliding window method was used to segment the acquired star surface images at different scales to obtain star surface image patches of different scales, such as... Figure 3 As shown, the star surface image patches at different scales contain impact crater targets with different diameter ranges.

[0070] After obtaining star catalog image patches at different scales, they are made into training and testing sets with no spatial intersection. Both the training and testing sets contain star catalog image patches at each scale.

[0071] In some embodiments of the present invention, step S1, which involves segmenting each star catalog image at different scales, specifically includes the following steps.

[0072] Three different sizes of sliding windows are set up: small-scale window, medium-scale window, and large-scale window. Specifically, in this embodiment, the dimensions of the three different sizes of sliding windows are 320×320 pixels, 640×640 pixels, and 1280×1280 pixels, respectively.

[0073] Three different sized sliding windows start from the top left corner of each star catalog image and slide sequentially from left to right and from top to bottom. During the sliding process, star catalog image blocks of the corresponding window size are cropped from the image, namely small-scale image blocks, medium-scale image blocks, and large-scale image blocks. The sizes of the image blocks of the three scales are 320×320 pixels, 640×640 pixels, and 1280×1280 pixels, respectively.

[0074] In some embodiments of the present invention, the following steps are also included.

[0075] A first diameter threshold and a second diameter threshold are set, wherein the second diameter threshold is greater than the first diameter threshold. In this embodiment, the first diameter threshold is 50 pixels and the second diameter threshold is 100 pixels.

[0076] The image patches cropped by the three different sizes of sliding windows respectively contain impact crater targets with diameters below the first diameter threshold, impact crater targets between the first diameter threshold and the second diameter threshold, and impact crater targets above the second diameter threshold.

[0077] Specifically, a 320×320 pixel image patch must contain at least impact crater targets with a diameter of less than 50 pixels; a 640×640 pixel image patch must contain at least impact crater targets with a diameter between 50 and 100 pixels; and a 1280×1280 pixel image patch must contain at least impact crater targets with a diameter of more than 100 pixels. It should be noted that, in addition to containing impact crater targets of the corresponding diameter, each size of image patch may also contain impact crater targets of other diameters.

[0078] In terms of image feature extraction, in order to solve the problem that existing technologies lack sufficient feature information when identifying small-scale targets, some embodiments of the present invention adopt a "slice-enlarge-slice" (SES) strategy for small-scale image blocks, which further includes the following steps.

[0079] A predetermined scale threshold is set. In this embodiment, the value of the predetermined scale threshold is between 320 pixels and 640 pixels.

[0080] In step S1, after obtaining star catalog image patches of different scales, for image patches with side lengths below a predetermined scale threshold, they are divided into multiple sub-regions of equal size. Each sub-region is then magnified to the same scale as the original image patch before magnification. Based on the division of each sub-region, the magnified image patch is further segmented into multiple image patches of the same scale as the original image patch before magnification. For example, see attached... Figure 4 As shown, in this embodiment, a "slice-enlarge-slice" strategy is adopted for the 320×320 pixel image block. The obtained 320×320 pixel image block is upsampled twice and enlarged into a 1280×1280 pixel image block. The enlarged image block is divided into 320×320 pixel regions and cropped into 16 320×320 pixel image blocks.

[0081] S2: Construct a multi-scale impact crater detection model, which includes three structurally identical and independent YOLOv5-based impact crater detection sub-models.

[0082] S3: The three impact crater detection sub-models are trained using the training set data to obtain three sub-models for detecting and recognizing impact crater targets with different diameter ranges. In this embodiment, they are the YOLOv5-L sub-model, the YOLOv5-M sub-model, and the YOLOv5-S sub-model, respectively. The three sub-models are responsible for detecting large-scale (above 100 pixels), medium-scale (50-100 pixels), and small-scale (below 50 pixels) impact crater targets in the region, respectively. Each sub-model adjusts its weights through training to optimize for impact craters of a specific scale.

[0083] After the sub-models are trained, the detection and recognition performance of the three sub-models is tested using test set data.

[0084] S4: Obtain the star catalog image to be detected, and divide it into image blocks of 320×320 pixels, 640×640 pixels, and 1280×1280 pixels respectively. Input these blocks into the YOLOv5-S sub-model, YOLOv5-M sub-model, and YOLOv5-L sub-model respectively to obtain all candidate bounding boxes for impact crater targets in the star catalog image to be detected, as shown in the attached figure. Figure 4 As shown. It should be noted that multiple candidate bounding boxes for the same impact crater target can be obtained by labeling a single sub-model, or by labeling multiple sub-models separately.

[0085] S5: Sort all candidate bounding boxes for each impact crater target based on their area, calculate the IoU value between the candidate bounding box with the largest area and other candidate bounding boxes, and select the candidate bounding box with the largest area as the final detection result of the candidate bounding box among all candidate bounding boxes whose IoU value exceeds the predetermined IoU threshold.

[0086] It should be noted that YOLOv5 uses the traditional non-maximum suppression (NMS) algorithm to select the predicted bounding box B1 with the highest confidence score from all candidate bounding boxes as the baseline, and then removes all other bounding boxes whose union intersection (IoU) with B1 exceeds a predetermined threshold. However, as shown in the attached... Figure 6 As shown, the best-detected target in the target region is not necessarily the target with the highest confidence score. Therefore, in some specific scenarios, the traditional NMS algorithm cannot select the optimal candidate bounding box.

[0087] This invention provides a region-first non-maximum suppression algorithm (NMS_AF), whose core idea is to consider not only the confidence level but also the area of ​​the candidate bounding boxes when handling IoU overlap. When the IoU value of multiple candidate bounding boxes with a candidate bounding box having the highest confidence level exceeds a predetermined IoU threshold, the candidate bounding boxes exceeding the threshold are grouped together, and the candidate bounding box with the largest area in this group is selected as the representative of the group, i.e., the final detection result for the impact crater target. In practical applications, the suppression threshold can be dynamically adjusted according to the candidate bounding box scores, overlap, and positional relationships to ensure that the algorithm selects the optimal candidate bounding box as the final detection result, reducing false positives and false negatives.

[0088] In some embodiments of the present invention, S5 specifically includes the following steps.

[0089] S51: Define the predetermined IoU threshold τ and the candidate bounding box list B for each impact crater target:

[0090] B = {B1, B2, ..., B} N};

[0091] Each candidate bounding box contains location information.

[0092] S52: Sort the candidate bounding boxes in list B based on their area.

[0093] S53: Select the candidate bounding box B with the largest current area from list B. i , i∈[1,N].

[0094] Calculate the other candidate bounding boxes in list B and B i The IoU value, which will include all values ​​related to B i Candidate bounding boxes whose IoU values ​​exceed a predetermined IoU threshold τ are added to a predetermined empty set G.

[0095] If set G is not empty after the calculation, then in B i Select the candidate bounding box B{max} with the largest area from set G and add it to the result list D.

[0096] S54: Remove the candidate bounding box with the largest current area from list B and all candidate bounding boxes from set G.

[0097] S55: Repeat steps S53-S54 until all candidate bounding boxes in list B are removed.

[0098] S56: Output the candidate bounding boxes in list D as the final detection result for the impact crater target.

[0099] This invention, by constructing a multi-scale impact crater detection model and introducing a region-first non-maximum suppression algorithm (NMS_AF), not only increases the accuracy and efficiency of target identification for impact craters of different scales, but also overcomes the identification difficulties caused by the scarcity of features and uneven sample distribution of small impact craters in the prior art by adopting a "slice-enlarge-slice" strategy.

[0100] The beneficial effects of the present invention are illustrated below with reference to a specific embodiment:

[0101] In a specific embodiment of identifying 500,000 impact craters in the Chang'e-4 landing area, the evaluation of the detection results of the multi-scale impact crater detection model and its sub-models is attached. Figure 8 As shown, attached Figure 8 The statistical data for the mid-level evaluation are shown in Table 1, and the impact crater recognition performance is shown in the attached figure. Figure 9 As shown. Based on the attached Figure 8 Appendix Figure 9As shown in Table 1, in terms of recognition accuracy, the detection method YOLOv5-Fused provided by this invention identified approximately 500,000 impact craters in the Chang'e-4 landing area. The smallest identifiable diameter of the impact craters can reach 4m. Furthermore, the multi-scale impact crater detection model has a Precision of 0.985, a Recall of 0.943, and an F1-Score of 0.964, demonstrating excellent performance.

[0102] Table 1 Evaluation of Detection Results of Multi-Scale Impact Crater Detection Model and its Sub-models

[0103]

[0104] It should be noted that in traditional impact crater detection methods, the values ​​of evaluation indicators such as precision, recall, and F1 score are usually between 0.85 and 0.90. However, the detection method provided by this invention improves each indicator by 5%-10%. This means that when identifying impact craters of different scales, especially small impact craters, the detection method provided by this invention can more accurately separate impact crater targets of different scales from complex backgrounds, reducing false alarms and false negatives.

[0105] Some embodiments of the present invention further provide a multi-scale star catalog impact crater detection system, characterized in that it includes:

[0106] At least one processor;

[0107] At least one memory for storing at least one program;

[0108] When at least one program is executed by at least one processor, the at least one processor implements the multi-scale star catalog impact crater detection method provided by the present invention.

[0109] Some embodiments of the present invention further provide a storage medium storing a processor-executable program, characterized in that: the processor-executable program, when executed by a processor, is used to implement a multi-scale star catalog impact crater detection method provided by the present invention.

[0110] Finally, it should be noted that the various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The same or similar parts between the various embodiments can be referred to each other.

[0111] The above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them; although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications can still be made to the specific implementation of the present invention or equivalent substitutions can be made to some technical features without departing from the spirit of the technical solutions of the present invention, and all such modifications and substitutions should be covered within the scope of the technical solutions claimed in the present invention.

Claims

1. A multi-scale catalog impact crater detection method, characterized in that, Includes the following steps: S1: Acquire star surface image data containing impact craters, and segment the acquired star surface images at different scales to obtain a dataset of star surface image patches at different scales; the star surface image patches at different scales contain impact crater targets with different diameter ranges respectively. Set a predetermined scale threshold; after obtaining star catalog image blocks of different scales, for image blocks with side lengths below the predetermined scale threshold, divide them into multiple sub-regions of the same size, enlarge each sub-region separately, and further divide the enlarged image blocks based on the division of each sub-region. S2: Construct a multi-scale impact crater detection model, which includes multiple impact crater detection sub-models based on YOLOv5; S3: Multiple impact crater detection sub-models are trained using datasets of star catalog image patches at different scales to obtain multiple sub-models for detecting and recognizing impact crater targets with different diameter ranges. S4: Obtain the star catalog image to be detected, divide it into image blocks of different scales, and input them into the sub-model of the corresponding scale range to obtain all candidate bounding boxes of each impact crater target in the star catalog image to be detected. S5: Sort all candidate bounding boxes for each impact crater target based on the area of ​​the candidate bounding boxes, calculate the IoU value between the candidate bounding box with the largest area and other candidate bounding boxes, and select the candidate bounding box with the largest area as the final detection result of the impact crater target among all candidate bounding boxes with IoU values ​​exceeding the predetermined IoU threshold. S5 specifically includes the following steps: S51: Define the predetermined IoU threshold τ and the candidate bounding box list B for each impact crater target: B = { B1, B2, …, BN}; Each candidate bounding box contains location information; S52: Sort the candidate bounding boxes in list B based on their area; S53: Select the candidate bounding box Bi with the largest current area in list B, i∈[1,N]; Calculate the IoU value between other candidate bounding boxes in list B and Bi, and add all candidate bounding boxes whose IoU value with Bi exceeds a predetermined IoU threshold τ to a predetermined empty set G; After the calculation is complete, if set G is not empty, select the candidate bounding box B{max} with the largest area from Bi and set G and add it to the result list D; S54: Remove the candidate bounding box with the largest current area from list B and all candidate bounding boxes from set G; S55: Repeat steps S53-S54 until all candidate bounding boxes in list B are removed; S56: Output the candidate bounding boxes in list D as the final detection result of the impact crater target.

2. The multi-scale star catalog impact crater detection method according to claim 1, characterized in that, Step S1, the method of segmenting each star catalog image at different scales specifically includes the following steps: Set up three different sized sliding windows; Three different sized sliding windows start from the top left corner of each star catalog image and slide sequentially from left to right and from top to bottom, cropping out star catalog image blocks of the corresponding window size from the image during the sliding process.

3. The multi-scale star catalog impact crater detection method according to claim 2, characterized in that, The three different sizes of sliding windows are 320×320 pixels, 640×640 pixels, and 1280×1280 pixels, respectively.

4. The multi-scale star catalog impact crater detection method according to any one of claims 2 or 3, characterized in that, It also includes the following steps: Set a first diameter threshold and a second diameter threshold; the second diameter threshold is greater than the first diameter threshold. The image patches cropped by the three different sizes of sliding windows respectively contain impact crater targets with diameters below the first diameter threshold, impact crater targets between the first diameter threshold and the second diameter threshold, and impact crater targets above the second diameter threshold.

5. The multi-scale star catalog impact crater detection method according to claim 4, characterized in that, The first diameter threshold is 50 pixels, and the second diameter threshold is 100 pixels.

6. The multi-scale star catalog impact crater detection method according to claim 1, characterized in that, The predetermined scale threshold ranges from 320 pixels to 640 pixels.

7. A multi-scale star catalog impact crater detection system, characterized in that, include: At least one processor; At least one memory for storing at least one program; When the at least one program is executed by the at least one processor, the at least one processor implements the multi-scale star catalog impact crater detection method as described in any one of claims 1-6.

8. A storage medium storing a processor-executable program, characterized in that: The processor-executable program, when executed by the processor, is used to implement the multi-scale star catalog impact crater detection method as described in any one of claims 1-6.

Images