Target structure registration measurement method and system for planet and asteroid exploration

Abstract

The present invention discloses a target structure registration measurement method and system for planet and asteroid exploration. The method is implemented by a time-of-flight camera installed on a satellite. The method comprises: receiving images and depth information of the surface environment of a target star body collected regularly by the time-of-flight camera, fusing the images and depth information, and generating three-dimensional point cloud data; filtering the three-dimensional point cloud data to obtain denoised three-dimensional point cloud data, and storing the denoised three-dimensional point cloud data in a data list; extracting effective area information from the denoised three-dimensional point cloud data to obtain data of the effective point cloud area; then performing iterative closest point method registration to obtain a transformation matrix; extracting the denoised three-dimensional point cloud from the data list, transforming it using the transformation matrix, and obtaining registered overall three-dimensional point cloud data; and extracting features from the registered overall three-dimensional point cloud data to obtain distance information, position information, and shape information of the surface of the target star body.

Description

Technical Field

[0001] The present invention relates to the field of data analysis, and in particular to a target structure registration measurement method and system for planet and asteroid exploration. Background Art

[0002] Planet and asteroid exploration can help us understand the origin and evolution of the solar system. Through exploration, we hope to study the material sources of the solar system and understand the fractionation, condensation, and formation processes of the solar nebula; study the characteristics, origin, and evolution of the atmospheres, ionospheres, and magnetic fields of planets and satellites; conduct comparative studies of the characteristics and evolutionary history of the topography, geological structure, and internal structure of planets and satellites; understand the utilization prospects of the resources, energy, and special environments of planets and satellites; and study and measure the time series of solar system evolution (element ages, formation interval ages, condensation ages of celestial bodies, solidification ages, metamorphic ages, gas retention ages, track retention ages, cosmic exposure ages, and landing ages) to infer the origin and evolution of the solar system.

[0003] With the rapid development of aerospace technology and the increasing attention paid by countries around the world to exploring the mysteries of the universe and developing and utilizing cosmic resources, human planetary exploration activities will become increasingly active, the purpose and content of exploration will become more in-depth and clear, the means will become more complete and advanced, and space exploration will enter a new stage of development.

[0004] With the continuous development of planet and asteroid exploration, we need to use various more advanced technologies to measure the relevant characteristics of various celestial bodies. However, the current development of measurement technology limits further research on celestial bodies, especially the measurement technology of local surface characteristics of celestial bodies, which affects whether we can better detect planets and asteroids. Summary of the Invention

[0005] The purpose of the present invention is to overcome the shortcomings of the prior art and propose a geometric target structure alignment measurement method and system for planet and asteroid exploration.

[0006] The present invention proposes a target structure registration measurement method for planet and asteroid exploration, which is implemented by a time-of-flight camera installed on a satellite. The method comprises:

[0007] Receive the image and depth information of the target star's surface environment collected by the time-of-flight camera, fuse the image and depth information, and generate three-dimensional point cloud data;

[0008] Perform filtering on the three-dimensional point cloud data to obtain denoised three-dimensional point cloud data, and store the denoised three-dimensional point cloud data in a data list;

[0009] Extract the effective area information from the denoised three-dimensional point cloud data to obtain the data of the effective point cloud area;

[0010] Perform iterative closest point registration on the data in the valid point cloud area to obtain the transformation matrix;

[0011] Extract the denoised 3D point cloud from the data list, transform it using the transformation matrix, and obtain the registered overall 3D point cloud data;

[0012] Feature extraction is performed on the registered overall three-dimensional point cloud data to obtain the distance information, position information and shape information of the target star surface.

[0013] As an improvement to the above method, the method of receiving the image and depth information of the target celestial body surface environment regularly acquired by the time-of-flight camera, fusing the image and depth information, and generating three-dimensional point cloud data specifically includes:

[0014] Receive images and depth information of the target celestial body's surface environment collected regularly by a time-of-flight camera;

[0015] Perform preliminary correction and temperature calibration on depth information;

[0016] Perform distortion correction on the image;

[0017] The image and depth information are converted from the depth image coordinate system to the camera coordinate system, and then to the world coordinate system to obtain three-dimensional point cloud data, where the data values ​​of the X-axis and Y-axis represent the plane coordinate position of the scene point, and the data value of the Z-axis represents the depth value of the scene point.

[0018] As an improvement to the above method, the filtering process is performed on the three-dimensional point cloud data to obtain denoised three-dimensional point cloud data, and the denoised three-dimensional point cloud data is stored in a data list; specifically, the filtering process includes:

[0019] Convert the 3D point cloud data into an M×N×3 matrix, where M×N represents the number of pixels in a frame of point cloud data.

[0020] Reset the M×N×3 matrix to an M×N matrix, and use the depth value to represent the value of each element in the reset matrix, thereby converting the three-dimensional point cloud data into two-dimensional plane image data;

[0021] A 3×3 spatial filtering operator based on a 3D point cloud is used to calculate the depth difference between the pixel and the central pixel according to the depth value of each pixel in the 2D plane image data.

[0022] The depth difference of each pixel is compared with the preset global threshold. When the depth difference is greater than the preset global threshold, the depth value measured at the pixel is a noise point, and the pixel is removed from the 3D point cloud data; otherwise, the pixel is retained to obtain the denoised 3D point cloud data and store it in the data list.

[0023] As an improvement to the above method, the effective area information is extracted from the denoised three-dimensional point cloud data to obtain data of the effective point cloud area; specifically:

[0024] The Sobel edge extraction operator is used to extract the edge of the two-dimensional plane image information to obtain edge information, and the points within the edge information are mapped three-dimensionally to obtain the data of the valid point cloud area and store it in the data list.

[0025] As an improvement to the above method, the data in the valid point cloud area is iteratively registered using the closest point method to obtain a transformation matrix; specifically:

[0026] Step 1) Take a group of point cloud data corresponding to the same valid area from the data list, set part of the point cloud data as the source point cloud, and set the other part as the target point cloud;

[0027] Step 2) Take the point set X1 from the target point cloud and find the point set X0 in the source point cloud that represents the same target based on the similar features of the point cloud data;

[0028] Step 3) Preliminary estimation is performed based on X1 and X0 to obtain the initial rotation matrix R0 and the initial translation matrix T0; the initial value of the iteration number i is 1;

[0029] Step 4) Calculate the i-th rotation matrix R according to the following formula i and the translation matrix T i :

[0030]

[0031] T i =pR i q

[0032] Among them, p j is a point in X0, q j is a point in X1, n is the number of points in X0, p=(p1,…,p j …,p n ),q=(q1,…,q j ,…q n );

[0033] Step 5) Use the rotation matrix R on X1 in turn i Perform rotation transformation using translation matrix T iPerform translation transformation to obtain a new corresponding point set X2, assign X2 to X1, and calculate the error delt between X1 and X0; determine whether delt is greater than a predetermined threshold. If so, after i+1, go to step 4); otherwise, go to step 6);

[0034] Step 6) The current rotation matrix R i and the translation matrix T i Combine into a transformation matrix.

[0035] As an improvement to the above method, feature extraction is performed on the registered overall three-dimensional point cloud data to obtain distance information, position information, and shape information of the target star surface; specifically, the feature extraction includes:

[0036] The Sobel edge extraction operator is used to extract the edge of the depth information of the overall 3D point cloud data after registration. Morphological processing is then used to extract the effective area information on the image. The point cloud of local features is obtained through mapping. The distance information, position information and shape information of the local features are then calculated and saved.

[0037] The present invention proposes a target structure registration and measurement system for planet and asteroid exploration, the system comprising: a time-of-flight camera, a preprocessing module, a denoising module, a region information extraction module, a transformation matrix generation module, a registration module, and a target result output module; wherein,

[0038] The time-of-flight camera is installed on the satellite and is used to regularly collect images and depth information of the surface environment of the target celestial body;

[0039] The pre-processing module is used to receive the image and depth information of the target star surface environment collected by the time-of-flight camera, fuse the image and depth information, and generate three-dimensional point cloud data;

[0040] The denoising processing module is used to perform filtering processing on the three-dimensional point cloud data to obtain denoised three-dimensional point cloud data, and store the denoised three-dimensional point cloud data in a data list;

[0041] The region information extraction module is used to extract effective region information from the denoised three-dimensional point cloud data to obtain data of the effective point cloud region;

[0042] The transformation matrix generation module is used to perform iterative closest point registration on the data of the valid point cloud area to obtain a transformation matrix;

[0043] The registration module is used to extract the denoised three-dimensional point cloud from the data list and transform it using the transformation matrix to obtain the registered overall three-dimensional point cloud data;

[0044] The target result output module is used to extract features from the registered overall three-dimensional point cloud data to obtain distance information, position information and shape information of the target star surface.

[0045] Compared with the prior art, the advantages of the present invention are:

[0046] 1. The method of the present invention can collect image data without being affected by external lighting, and can collect image data of the surface environment of a celestial body even when the celestial body is dark;

[0047] 2. The method of the present invention uses target registration to achieve the fusion of point cloud data collected multiple times, thereby improving the accuracy of target structure measurement. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 This is a flow chart of a target structure registration measurement method for planet and asteroid exploration according to embodiment 1 of the present invention;

[0049] Figure 2 This is a composition diagram of a target structure alignment and measurement system for planet and asteroid exploration according to embodiment 2 of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] 201. Time-of-flight camera 202. Preprocessing module

[0052] 203. Denoising processing module 204. Regional information extraction module

[0053] 205. Transformation matrix generation module 206. Registration module

[0054] 207. Target result output module DETAILED DESCRIPTION

[0055] In response to the problems existing in the prior art, the purpose of the embodiments of the present invention is to provide a method and system for geometric target structure registration measurement for planet and asteroid exploration. By using a Time of Flight (TOF) camera, environmental images of the surface of planets and asteroids are collected at a preset frequency, and three-dimensional point cloud data is generated and sent to a satellite processor. The collection of image data is not affected by external lighting, and image data of the surface environment of the planet can be collected even when the planet is dark. The satellite processor analyzes the multiple received three-dimensional point cloud data to obtain surface information of the object, and then extracts the effective area of ​​the surface information of the multiple point clouds, thereby registering the effective areas of the multiple point clouds, thereby achieving registration of the entire area. Finally, the registered point cloud data is used for calculation to obtain the corresponding data.

[0056] The present invention provides a method and system for geometric target structure registration measurement for planet and asteroid exploration. Using a time-of-flight (TOF) camera, the system captures environmental images of the planetary and asteroidal surfaces at a preset frequency, generating three-dimensional point cloud data that is sent to a satellite processor. This image data is not affected by external illumination, allowing the capture of surface environmental image data even when the planet is dark. The satellite processor analyzes the received 3D point cloud data to obtain surface information of the object. It then extracts valid regions from the surface information of the multiple point clouds, registering the valid regions of the multiple point clouds and achieving registration of the entire region. Finally, the registered point cloud data is used for calculations to obtain the corresponding data.

[0057] The technical solution of the present invention is described in detail below with reference to the accompanying drawings and embodiments.

[0058] Example 1

[0059] like Figure 1 As shown, Example 1 of the present invention provides a geometric target structure registration and measurement method for planetary and asteroid exploration. By using a TOF camera to capture images of satellites and asteroid regions, three-dimensional point cloud data is generated. Surface structure information of the planets and asteroids is obtained by analyzing and calculating the three-dimensional point cloud data. This geometric target structure registration and measurement method for planetary and asteroid exploration, provided in this embodiment of the present invention, achieves fully automated extraction of planetary and asteroid surface structure information without being affected by ambient light.

[0060] The specific steps include:

[0061] In step 110, a time-of-flight (TOF) camera captures environmental image information of the monitoring area of ​​the bridge according to an image acquisition instruction. The TOF camera's processing unit analyzes and processes the collected image environmental information to generate three-dimensional point cloud data.

[0062] The three-dimensional point cloud data includes three-dimensional point cloud data of the marked graphics, and the three-dimensional point cloud data has corresponding acquisition time data; the TOF camera has a camera ID.

[0063] Specifically, the TOF camera receives image acquisition commands from the satellite processor, captures multiple frames of environmental images of the planetary and asteroidal regions, and generates 3D point cloud data. As the TOF camera captures the environmental images of the region, it saves the time data associated with the image acquisition. Multiple captures of the surface structures of the planets and asteroids provide more data, allowing for the extraction of more detailed information on surface structural features. This allows for richer data collection that meets specific requirements.

[0064] The TOF camera is mounted on a satellite. Since the satellite is in constant motion, it can obtain a large amount of data information from different areas, making it convenient to obtain various structural information on the surface of the entire planet and asteroid.

[0065] The TOF camera used in this embodiment transmits light signals through a built-in laser emission module, and acquires the distance and depth of field data of the three-dimensional scene through a built-in complementary metal oxide semiconductor (CMOS) pixel array. The imaging rate can reach hundreds of frames per second, while the structure is compact and the power consumption is low. The three-dimensional data acquisition method for the target scene is as follows: the TOF camera uses an amplitude-modulated light source, which actively illuminates the target scene and is coupled with the relevant sensor on each pixel locked at the same frequency. The emitted light emitted by its built-in laser and the reflected light emitted after the emitted light hits the scene object have a phase shift, and multiple measurements are obtained by detecting the different phase shifts between the emitted light and the reflected light. The amplitude modulation of the built-in laser emitter is in the modulation frequency range of 10-100MHz, and the frequency controls the depth range and depth resolution of the TOF camera sensor. At the same time, the processing unit of the TOF camera performs phase difference calculation on each pixel separately to obtain the depth data of the target scene. The processing unit of the TOF camera analyzes and calculates the reflection intensity of the reflected light to obtain the intensity data of the target scene, and then combines the acquired two-dimensional data for analysis and processing to obtain the three-dimensional point cloud data of the target scene.

[0066] In this embodiment, the TOF camera uses a solid-state laser or LED array to emit a light wave transmitter with a wavelength of approximately 850nm as a built-in laser transmitter. The emitted light source is a continuous square wave or sine wave obtained through continuous modulation. The TOF camera processing unit calculates the phase angle of the emitted and reflected light in multiple samples and the distance to the target object, and analyzes and calculates the current intensity converted from the reflected light intensity to obtain intensity data. This data is then combined with the two-dimensional image data obtained by the optical camera for fusion processing to obtain three-dimensional point cloud data of the target scene.

[0067] In the process of capturing the target area's environmental image, the TOF camera actively emits non-visible light to capture the scene, so even in dark conditions, clear 3D point cloud data of the target area's environmental image can be obtained. Therefore, this method is also applicable in dark environments.

[0068] In a preferred embodiment of this embodiment, before the TOF camera captures an image of the target area according to the image acquisition instruction, the satellite processor receives a start command and generates an image acquisition instruction based on a preset time interval; the satellite processor then sends the image acquisition instruction to the TOF camera. In other words, when the method provided by the embodiment of the present invention needs to be initiated, an administrator enters a start command on the satellite processor's interactive screen or sends the start command to the satellite processor through other means. After receiving the start command, the satellite processor reads the preset time interval and generates an image acquisition instruction based on the time interval. For example, if the time interval is 5 minutes, the satellite processor will generate an image acquisition instruction every 5 minutes and send it to the TOF camera.

[0069] In step 120 , the TOF camera sends the three-dimensional point cloud data to the satellite for processing.

[0070] Specifically, the TOF camera sends the collected and generated three-dimensional point cloud data to the satellite for processing.

[0071] Step 130 , performing denoising processing on the three-dimensional point cloud data to obtain denoised three-dimensional point cloud data.

[0072] Specifically, a specific filtering method is selected to filter the received 3D point cloud data to remove noise. For example, in this embodiment, the following method is used to filter the 3D point cloud data:

[0073] The TOF camera resolution in this embodiment is M×N (M and N are both positive integers), so the acquired three-dimensional point cloud data has M×N pixels, and each pixel further includes X, Y, and Z three-dimensional coordinate values. The steps for converting the TOF camera's raw depth data to the three-dimensional point cloud data we need are: first, preliminary correction and temperature calibration are performed on the raw depth data; second, distortion correction is performed on the image; third, the depth image coordinate system (x0, y0, z0) is converted to the camera coordinate system (x1, y1, z1), that is, the depth information on the image is converted into a three-dimensional coordinate system with the camera as the origin; finally, the camera coordinate system (x1, y1, z1) is converted to the required world coordinate system (x2, y2, z2), that is, the camera coordinate system is converted into the coordinate system required by the project, which is the coordinate system of the final point cloud. The data values ​​of the X-axis and Y-axis represent the plane coordinate position of the scene point, and the data value of the Z-axis represents the actual depth value of the captured scene.

[0074] The 3D point cloud data is converted into an M×N×3 matrix, where each row vector represents a pixel arranged in the time-of-flight sensor. By reshaping the M×N×3 matrix into an M×N matrix and representing the value of each element in the reshaping matrix with a depth value, the 3D point cloud data is converted into 2D planar image data.

[0075] A 3×3 spatial filtering operator based on the 3D point cloud is used to calculate the depth value of each pixel in the 2D image data. The depth difference between the center pixel and the surrounding pixels is calculated. This depth difference is compared with a preset global threshold. If the depth difference exceeds the threshold, the depth value measured at that pixel is considered noise, and the corresponding pixel in the 3D point cloud data is filtered out. Otherwise, the corresponding pixel in the 3D point cloud data is retained. After processing, denoised 3D point cloud data is obtained.

[0076] Step 140 : extracting effective area information from the denoised three-dimensional point cloud data to obtain data of the effective point cloud area.

[0077] Specifically, a specific extraction method is selected to extract valid areas from the received 3D point cloud data and remove invalid information. For example, in this embodiment, the following method is used to extract valid area information from the 3D point cloud data:

[0078] The satellite processor uses the Sobel edge extraction operator to extract edges from the 2D plane image information of the 3D point cloud data. Points within the edge information of the 2D plane image are mapped 3D to extract the valid area information of the 3D point cloud.

[0079] Step 150 : Perform iterative closest point registration on the valid point cloud area to obtain a transformation matrix and registered data.

[0080] The satellite processor selects a specific iterative closest point method to perform a registration operation on the received three-dimensional point cloud data to obtain a registration matrix. For example, in this embodiment, the following method is used to perform an iterative closest point operation on the three-dimensional point cloud data:

[0081] A set of corresponding point cloud data is defined as target point cloud and source point cloud respectively.

[0082] Take a point set in the target point cloud and find the corresponding point set data in the source point cloud. Calculate the rotation matrix and translation matrix to minimize the error. Use the obtained rotation matrix and translation matrix to rotate and translate the target point cloud to obtain the new corresponding point set.

[0083] Repeat the above operation until the rotation matrix and translation matrix that meet the required accuracy requirements are obtained, and then merge the rotation matrix and translation matrix into the transformation matrix.

[0084] Specifically:

[0085] 1) Take out a group of point cloud data corresponding to the same valid area from the data list, set part of the point cloud data as the source point cloud, and set the other part as the target point cloud;

[0086] 2) Take the point set X1 from the target point cloud and find the point set X0 in the source point cloud that represents the same target based on the similar features of the point cloud data;

[0087] 3) The initial rotation matrix R0 and the initial translation matrix T0 are preliminarily estimated from X1 and X0; the initial value of the iteration number i is 1;

[0088] 4) Calculate the i-th rotation matrix R according to the following formula i and the translation matrix T i :

[0089]

[0090] T i =pR i q

[0091] Among them, p j is a point in X0, q j is a point in X1, n is the number of points in X0, p=(p1,…,p j …,p n ),q=(q1,…,q j ,…q n );

[0092] 5) Use the rotation matrix R for X1 in turn i Perform rotation transformation using translation matrix T i Perform a translation transformation to obtain a new corresponding point set X2, assign X2 to X1, and calculate the error delt between X1 and X0; determine whether delt is greater than a predetermined threshold. If so, add i+1 and go to 4); otherwise, go to 6);

[0093] 6) The current rotation matrix R i and the translation matrix T i Combine into a transformation matrix.

[0094] Step 160 : transforming the denoised three-dimensional data using a transformation matrix to obtain registered overall three-dimensional point cloud data.

[0095] Specifically, the transformation matrix is ​​used to perform transformation to obtain the registered overall three-dimensional point cloud data.

[0096] Step 170 : performing feature extraction on the registered overall three-dimensional point cloud data to obtain distance information, position information, and shape information of local features of multiple planets and asteroids, and save the information in a feature data list.

[0097] Specifically, the depth information of the three-dimensional point cloud data is used, the Sobel edge extraction operator is adopted, the depth information is used to extract the edge, and then morphological processing is used to extract the effective area information on the image. The point cloud of the local features is obtained by mapping, and the distance information, position information and shape information of the local features are calculated and saved in the feature data list.

[0098] This embodiment provides a geometric target structure registration measurement method for planet and asteroid exploration. Using a time-of-flight camera to capture images of satellites and asteroid regions, the method generates three-dimensional point cloud data. By analyzing and calculating this three-dimensional point cloud data, the surface structure information of the planets and asteroids is obtained. This measurement method achieves fully automated extraction of planet and asteroid surface structure information, unaffected by ambient light.

[0099] Example 2

[0100] Based on the above method, Figure 2 As shown, embodiment 2 of the present invention proposes a target structure registration measurement system for planet and asteroid exploration. The system includes: a time-of-flight camera 201, a preprocessing module 202, a denoising module 203, a region information extraction module 204, a transformation matrix generation module 205, a registration module 206, and a target result output module 207; wherein,

[0101] A time-of-flight camera 201 is installed on the satellite and is used to regularly collect images and depth information of the target celestial body's surface environment;

[0102] A pre-processing module 202 is used to fuse the image and depth information of the target star surface environment collected by the time-of-flight camera to generate three-dimensional point cloud data;

[0103] The denoising processing module 203 is used to perform filtering processing on the three-dimensional point cloud data to obtain denoised three-dimensional point cloud data, and store the denoised three-dimensional point cloud data in a data list;

[0104] The region information extraction module 204 is used to extract effective region information from the denoised three-dimensional point cloud data to obtain data of effective point cloud regions;

[0105] The transformation matrix generation module 205 is used to perform iterative closest point registration on the data of the valid point cloud area to obtain a transformation matrix;

[0106] A registration module 206 is used to extract the denoised three-dimensional point cloud from the data list and transform it using a transformation matrix to obtain the registered overall three-dimensional point cloud data;

[0107] The target result output module 207 is used to extract features from the registered overall three-dimensional point cloud data to obtain the distance information, position information and shape information of the target star surface.

[0108] Professionals should also be further aware that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the above description has generally described the components and steps of each example according to their functions. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0109] The steps of the methods or algorithms described in conjunction with the embodiments disclosed herein may be implemented using hardware, a software module executed by a processor, or a combination of the two. The software module may be placed in a random access memory (RAM), a memory, a read-only memory (ROM), an electrically programmable ROM, an electrically erasable programmable ROM, a register, a hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art.

[0110] Finally, it should be noted that the above embodiments are intended only to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the embodiments, it should be understood by those skilled in the art that modifications or equivalent substitutions to the technical solutions of the present invention do not depart from the spirit and scope of the technical solutions of the present invention and are intended to be encompassed by the claims of the present invention.

Claims

1. A target structure registration measurement method for planet and asteroid exploration, implemented using a time-of-flight camera mounted on a satellite, comprising: Receive the image and depth information of the target star's surface environment collected by the time-of-flight camera, fuse the image and depth information, and generate three-dimensional point cloud data; Perform filtering on the three-dimensional point cloud data to obtain denoised three-dimensional point cloud data, and store the denoised three-dimensional point cloud data in a data list; Extract the effective area information from the denoised three-dimensional point cloud data to obtain the data of the effective point cloud area; Perform iterative closest point registration on the data in the valid point cloud area to obtain the transformation matrix; Extract the denoised 3D point cloud from the data list, transform it using the transformation matrix, and obtain the registered overall 3D point cloud data; Perform feature extraction on the registered overall 3D point cloud data to obtain the distance information, position information and shape information of the target star surface; The data of the valid point cloud area are aligned using an iterative closest point method to obtain a transformation matrix; Specifically: Step 1) Take a group of point cloud data corresponding to the same valid area from the data list, set part of the point cloud data as the source point cloud, and set the other part as the target point cloud; Step 2) Take the point set X1 from the target point cloud and find the point set X0 in the source point cloud that represents the same target based on the similar features of the point cloud data; Step 3) Preliminary estimation is performed based on X1 and X0 to obtain the initial rotation matrix R0 and the initial translation matrix T0; the initial value of the iteration number i is 1; Step 4) Calculate the i-th rotation matrix R according to the following formula i and the translation matrix T i : T i =p-R i q Among them, p j is a point in X0, q j is a point in X1, n is the number of points in X0, p=(p1,…,p j …,p n ),q=(q1,…,q j ,…q n ); Step 5) Use the rotation matrix R on X1 in turn i Perform rotation transformation using translation matrix T i Perform translation transformation to obtain a new corresponding point set X2, assign X2 to X1, and calculate the error delt between X1 and X0; determine whether delt is greater than a predetermined threshold. If so, after i+1, go to step 4); Otherwise go to step 6); Step 6) The current rotation matrix R i and the translation matrix T i Combine into a transformation matrix.

2. The target structure registration measurement method for planet and asteroid exploration according to claim 1, characterized in that: The method comprises: receiving the image and depth information of the target star surface environment collected by the time-of-flight camera at regular intervals, fusing the image and depth information, and generating three-dimensional point cloud data; specifically comprising: Receive images and depth information of the target celestial body's surface environment collected regularly by a time-of-flight camera; Perform preliminary correction and temperature calibration on depth information; Perform distortion correction on the image; The image and depth information are converted from the depth image coordinate system to the camera coordinate system, and then to the world coordinate system to obtain three-dimensional point cloud data, where the data values ​​of the X-axis and Y-axis represent the plane coordinate position of the scene point, and the data value of the Z-axis represents the depth value of the scene point.

3. The target structure registration measurement method for planet and asteroid exploration according to claim 2, characterized in that: The filtering process is performed on the three-dimensional point cloud data to obtain denoised three-dimensional point cloud data, and the denoised three-dimensional point cloud data is stored in a data list; specifically comprising: Convert the 3D point cloud data into an M×N×3 matrix, where M×N represents the number of pixels in a frame of point cloud data. Reset the M×N×3 matrix to an M×N matrix, and use the depth value to represent the value of each element in the reset matrix, thereby converting the three-dimensional point cloud data into two-dimensional plane image data; A 3×3 spatial filtering operator based on a 3D point cloud is used to calculate the depth difference between the pixel and the central pixel according to the depth value of each pixel in the 2D plane image data. The depth difference of each pixel is compared with the preset global threshold. When the depth difference is greater than the preset global threshold, the depth value measured at the pixel is a noise point, and the pixel is removed from the 3D point cloud data; otherwise, the pixel is retained to obtain the denoised 3D point cloud data and store it in the data list.

4. The target structure registration measurement method for planet and asteroid exploration according to claim 3, characterized in that: The extraction of effective area information from the denoised three-dimensional point cloud data to obtain data of the effective point cloud area is specifically as follows: The Sobel edge extraction operator is used to extract the edge of the two-dimensional plane image information to obtain edge information, and the points within the edge information are mapped three-dimensionally to obtain the data of the valid point cloud area and store it in the data list.

5. The target structure registration measurement method for planet and asteroid exploration according to claim 4, characterized in that: The feature extraction of the registered overall three-dimensional point cloud data to obtain the distance information, position information and shape information of the target star surface specifically includes: The Sobel edge extraction operator is used to extract the edge of the depth information of the overall 3D point cloud data after registration. Morphological processing is then used to extract the effective area information on the image. The point cloud of local features is obtained through mapping. The distance information, position information and shape information of the local features are then calculated and saved.

6. A target structure registration and measurement system for planet and asteroid exploration, characterized in that: The system includes: a time-of-flight camera, a pre-processing module, a denoising processing module, a region information extraction module, a transformation matrix generation module, a registration module and a target result output module; wherein, The time-of-flight camera is installed on the satellite and is used to regularly collect images and depth information of the surface environment of the target celestial body; The pre-processing module is used to receive the image and depth information of the target star surface environment collected by the time-of-flight camera, fuse the image and depth information, and generate three-dimensional point cloud data; The denoising processing module is used to perform filtering processing on the three-dimensional point cloud data to obtain denoised three-dimensional point cloud data, and store the denoised three-dimensional point cloud data in a data list; The region information extraction module is used to extract effective region information from the denoised three-dimensional point cloud data to obtain data of the effective point cloud region; The transformation matrix generation module is used to perform iterative closest point registration on the data of the valid point cloud area to obtain a transformation matrix; The registration module is used to extract the denoised three-dimensional point cloud from the data list and transform it using the transformation matrix to obtain the registered overall three-dimensional point cloud data; The target result output module is used to extract features from the registered overall three-dimensional point cloud data to obtain distance information, position information and shape information of the target star surface; The processing process of the transformation matrix generation module includes: Step 1) Take a group of point cloud data corresponding to the same valid area from the data list, set part of the point cloud data as the source point cloud, and set the other part as the target point cloud; Step 2) Take the point set X1 from the target point cloud and find the point set X0 in the source point cloud that represents the same target based on the similar features of the point cloud data; Step 3) Preliminary estimation is performed based on X1 and X0 to obtain the initial rotation matrix R0 and the initial translation matrix T0; the initial value of the iteration number i is 1; Step 4) Calculate the i-th rotation matrix R according to the following formula i and the translation matrix T i : T i =p-R i q Among them, p j is a point in X0, q j is a point in X1, n is the number of points in X0, p=(p1,…,p j …,p n ),q=(q1,…,q j ,…q n ); Step 5) Use the rotation matrix R on X1 in turn i Perform rotation transformation using translation matrix T i Perform translation transformation to obtain a new corresponding point set X2, assign X2 to X1, and calculate the error delt between X1 and X0; determine whether delt is greater than a predetermined threshold. If so, after i+1, go to step 4); otherwise, go to step 6); Step 6) The current rotation matrix R i and the translation matrix T i Combine into a transformation matrix.

Images