An occluded object reasoning search method based on a grid discrete representation accommodatability analysis method
By using mesh discrete representation and physical constraint analysis, the problem of searching for occluded objects in robot target search is solved, achieving efficient and accurate target localization, applicable to various container environments, and reducing computational complexity and data dependence.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- ROBOTICS RESEARCH CENTER OF YUYAO CITY
- Filing Date
- 2025-03-17
- Publication Date
- 2026-08-04
AI Technical Summary
Existing robot target search methods suffer from poor generalization when faced with occluded objects, high computational complexity, and failure to consider object placement constraints, leading to increased search difficulty and insufficient real-time performance.
A capacity analysis method based on grid discrete representation is adopted. Through image acquisition, object segmentation, grid discretization, calculation of the visible object's occupied area and capacity analysis, combined with support surface and height constraints, the potential location of the target object is inferred.
It improves the versatility and computational efficiency of the method, meets real-time requirements, has strong accuracy and robustness, does not require a large amount of labeled data, and the reasoning process is clear and transparent.
Smart Images
Figure CN120206512B_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of robot target search technology, and particularly relates to a reasoning-based search method for occluded objects based on the compatibility analysis of grid discrete representation. Background Technology
[0002] In robot target search tasks, target objects may be occluded by other objects, increasing the difficulty of the search. Existing target search methods are mainly divided into two categories: learning-based occupancy probability distribution methods and methods based on 3D information completion. However, these methods have the following problems:
[0003] Poor generalization: Learning-based methods require building a large number of labeled datasets for specific scenarios. When the scenario changes (such as object types, container types, lighting conditions, etc.), the model needs to be retrained, resulting in poor generalization ability.
[0004] High computational complexity: Methods based on 3D information completion require complex 3D reconstruction and optimization, resulting in a large amount of computation and making it difficult to meet real-time requirements.
[0005] Failure to consider object placement constraints: Existing methods rarely consider the physical constraints of object placement in the real world (such as support surfaces, height restrictions, etc.), which may lead to the inferred potential locations not conforming to reality. Summary of the Invention
[0006] The purpose of this invention is to provide a reasoning-based search method for occluded objects based on the compatibility analysis of grid discrete representation, so as to solve the above-mentioned technical problems.
[0007] To address the aforementioned technical problems, the present invention provides a specific technical solution for a reasoning-based search method for occluded objects based on a grid discrete representation capacity analysis method, as follows:
[0008] A reasoning-based search method for occluded objects based on grid discrete representation containment analysis includes the following steps:
[0009] Step 1: Image Acquisition and Preprocessing: The robot uses a camera to acquire RGB images of the inside of the container that may hold the object, and then performs preprocessing.
[0010] Step 2: Visual object detection and segmentation: Use image segmentation algorithms to segment visible objects in the image and identify their categories;
[0011] Step 3: Mesh Discretization: When an object is placed, the contact surface between the object and other objects is fixed, i.e., the supporting surface; when an object is placed inside a container, the surface supporting the object is usually fixed, i.e., the bearing surface; the bearing surface of the openable container and the supporting surface of the target object are projected onto a two-dimensional plane and discretized using a grid of set size, with each grid representing a small region.
[0012] Step 4: Calculate the area occupied by visible objects: Based on the segmentation results, calculate the area occupied by each visible object in the grid, mark the grid occupied by the container support surface as occupied, and the unoccupied grid as candidate positions; convert the support surface size of the target object into the grid size.
[0013] Step 5: Capacity Analysis: Perform a capacity analysis for each unoccupied grid; Step 6: Potential Location Ranking: Based on the results of the capacity analysis, sort all unoccupied grids;
[0014] Step 7: Reasoning Search: Examine each potential location sequentially according to the sorting results;
[0015] Step 8: Target Object Localization: The search ends when a target object is found at a potential location. Further, Step 2 includes extracting 3D point cloud information of visible objects by combining depth images. After obtaining the point cloud of visible objects in the scene, bounding box fitting can be used to obtain approximate structure, position, and pose information of the visible objects.
[0016] Furthermore, the capacity analysis in step 5 filters candidate containers in two stages: first, selecting openable containers that may store the target object; second, after opening the openable container, determining which visible objects might be blocking the target object's space. Using OBB to represent the target object to be searched, the capacity analysis problem is simplified to whether a cuboid of a given size can fit a polyhedron. The containers in the openable container selection stage are represented by cuboids; the containers in the visible object selection stage are represented by truncated cones. When an object is placed, the contact surface between the object and other objects is fixed, i.e., the supporting surface. When an object is placed inside a container, the surface supporting the placed object is usually fixed, i.e., the bearing surface. Therefore, the capacity problem of a target object in a specific container is transformed into the capacity problem of the object's height relative to the bearing surface and the capacity problem of the target object's supporting surface within the container's bottom surface.
[0017] Furthermore, step 5 includes the following steps:
[0018] Step 5.1: Compatibility analysis based on the support surface:
[0019] Step 5.2: Capacity analysis based on height constraints;
[0020] Step 5.3: Compatibility analysis combining support surface and height constraints:
[0021] Furthermore, in step 5.1, for openable containers, based on the gridded discrete representation, the number of consecutive unoccupied grids in the X and Y axes of the container's bearing surface area is detected to determine whether there exists a region where the number of consecutive unoccupied grids is greater than or equal to the size of the target object's grid. A two-dimensional rectangle is used to describe the bearing surface of the openable container. The condition for the target object to be accommodated in the openable container is that the height of the target object is lower than the height of the internal space of the openable container, and the supporting surface of the target object's bounding box can be accommodated within the rectangle of the openable container's bearing surface. If this exists, it is determined that the supporting surface of the target object can be accommodated in the container's bearing surface. If it cannot be accommodated in the current direction, the container's grid surface is rotated by a certain angle, and the above detection process is repeated until the cumulative rotation angle reaches a preset threshold. If there is still no accommodating solution, it is determined that the target object cannot be accommodated in the container.
[0022] Furthermore, step 5.1 uses a meshing method to discretize the target object and the bottom geometry of the container into a mesh, and analyzes the containment capacity between the two based on the mesh representation. The steps are as follows: Step 5.1.1: Construct a two-dimensional plane using the bottom information of the container, set the mesh size, and convert the plane into a mesh surface;
[0023] Step 5.1.2: Traverse all grids, set the grid value occupied by the container's carrying surface to 1, otherwise set it to 0, to determine the grid area that can be used to place the target item;
[0024] Step 5.1.3: Convert the support surface size of the target object into a grid number size using the grid size; detect the number of consecutive grids in the x and y directions of the container's bearing surface area. If the number of consecutive grids at a certain position is greater than the grid number size of the target object, then the support surface of the target object can be accommodated on the bottom surface of the container.
[0025] Step 5.1.4: If the target object's support surface cannot be accommodated on the bottom surface of the container in this analysis, rotate the container's grid surface by a certain angle and repeat steps 5.1.2 and 5.1.3. When the cumulative rotation angle reaches 90° and there is still no solution that can accommodate the target object, the target object cannot be accommodated in the container.
[0026] Furthermore, in step 5.2, for the occlusion space formed by the visible object, based on the principle of perspective projection, the occlusion space formed by the visible object on the supporting surface is calculated; it is determined whether the height of the target object is lower than the height of the occlusion space at that position when it is located at a certain point in the occlusion space; for areas that meet the height limit, the accommodability analysis based on the supporting surface described in step 5.1 is performed to obtain the possible positions of the target object in the occlusion space.
[0027] Furthermore, step 5.2 includes the following steps:
[0028] Assuming the viewpoint's position in the scene is l, the equation of the projection plane is defined as follows: Let p be the coordinates of a point on a visible object. The relationship between point p and its projection p' on the projection plane can be expressed as: p' = M·p
[0029] Where matrix M is: After obtaining the projection of the visible object onto the projection plane, the corresponding occlusion space is calculated.
[0030] To determine whether the height of a target object located at a certain point in the occlusion space is lower than the height of the occlusion space at that location: First, generate the shadow cast by the visible object on the supporting surface. Then, let the shadow area S... shadow The height of the interior point relative to the supporting surface is the height of the target object, resulting in a new point cloud S. shadow If S' shadow The projection of a point in the middle under this viewpoint is still in S. shadow If the point is inside, then the point is in S. shadow The corresponding point may accommodate the target item; for areas that meet the height restrictions, try in-plane accommodability analysis to obtain the possible locations of the target item in the scene.
[0031] Furthermore, step 5.3 comprehensively utilizes the capacity analysis methods described in steps 5.1 and 5.2 to decompose the capacity problem of the target object in a specific container into the capacity problem of the object's height relative to the bearing surface and the capacity problem of the target object's supporting surface in the bottom surface of the container. By combining the two, the possibility and potential location of the target object are determined.
[0032] Furthermore, the sorting in step 6 is based on the stability of the target object placed on the grid and the positional relationship between the grid and other known objects.
[0033] The occluded object reasoning search method based on grid discrete representation capacity analysis of the present invention has the following advantages:
[0034] 1. High versatility and wide applicability: This method does not depend on datasets for specific scenarios, does not require model training for different scenarios, and is applicable to various openable container environments, demonstrating strong versatility and adaptability.
[0035] 2. High computational efficiency and good real-time performance: This method mainly performs gridded discrete representation and simple geometric calculations on a two-dimensional plane, avoiding complex three-dimensional reconstruction and optimization processes, greatly reducing the amount of computation, and meeting the needs of robot real-time decision-making.
[0036] 3. High accuracy and robustness: This method fully considers the physical constraints of object placement in the real world, namely the support surface and height restrictions, so that the inferred potential position of the target object is more in line with the actual situation and avoids unreasonable prediction results.
[0037] 4. No need for a large amount of labeled data: This method does not require any training data, avoiding the cost and time of labeling data, and can be quickly deployed and applied.
[0038] 5. High interpretability: This method is based on clear geometric rules and physical constraints, and the reasoning process is clear and transparent, making it easy to understand and debug. Attached Figure Description
[0039] Figure 1 : Schematic diagram of two containment detection scenarios.
[0040] Figure 2 : Schematic diagram of capacity testing.
[0041] Figure 3 : Schematic diagram of containment detection based on gridded discrete method.
[0042] Figure 4 : Schematic diagram of the formation of the shading space.
[0043] Figure 5 : Schematic diagram of the accommodable area extraction based on height restrictions. Detailed Implementation
[0044] To better understand the purpose, structure, and function of this invention, the following description, in conjunction with the accompanying drawings, provides a more detailed explanation of the occluded object reasoning search method based on grid discrete representation containment analysis.
[0045] This invention proposes a reasoning-based search method for occluded objects based on the capacity analysis of grid discrete representation, specifically including the following steps:
[0046] Step 1: Image Acquisition and Preprocessing: The robot uses a camera to acquire RGB images of the inside of the container that may hold the object, and performs preprocessing, such as image denoising and color correction.
[0047] Step 2: Visual Object Detection and Segmentation: Image segmentation algorithms (e.g., Mask R-CNN) are used to segment visible objects (including partially visible objects) in the image and identify their categories. Combined with depth images, 3D point cloud information of visible objects is extracted. After obtaining the point cloud of visible objects in the scene, bounding box fitting is used to obtain approximate structure, position, and pose information of the visible objects.
[0048] Step 3: Grid Discrete Representation: such as Figure 1 As shown. In real life, when objects are placed, the contact surface between the object and other objects is fixed, i.e., the supporting surface. When objects are placed inside a container, the surface supporting the placed object is usually fixed, i.e., the bearing surface. The bearing surface of the openable container and the supporting surface of the target object are projected onto a two-dimensional plane and discretized using a grid of set size, with each grid representing a small region.
[0049] Step 4: Calculate the area occupied by visible objects: Based on the segmentation results, calculate the area occupied by each visible object in the mesh. Mark the mesh occupied by the container's supporting surface as occupied, and the unoccupied mesh as candidate positions; convert the dimensions of the target object's supporting surface into mesh size.
[0050] Step 5: Capacity Analysis: For each unoccupied mesh, perform a capacity analysis. This analysis filters candidate containers in two phases. First, it selects openable containers that might store the target object. Second, after opening the openable container, it determines which visible objects might occlude the target object within that space. Using OBB to represent the target object, the capacity analysis problem simplifies to whether a cuboid of a given size can fit into a polyhedron. Containers in the openable container selection phase are represented by cuboids; containers in the visible object selection phase are represented by truncated pyramids. The capacity problem of a target object within a specific container is transformed into the capacity problem of the object's height relative to its supporting surface and the capacity problem of the target object's supporting surface within the container's bottom surface.
[0051] Step 5.1: Compatibility Analysis Based on Support Surface: For openable containers, based on the aforementioned gridded discrete representation, the number of consecutive unoccupied grids in the X and Y axes of the container's support surface area is detected to determine whether there exists a region where the number of consecutive unoccupied grids is greater than or equal to the size of the target object's grid. A two-dimensional rectangle is used to describe the support surface of the openable container. The condition for the target object to be accommodated in the openable container is that the height of the target object is lower than the height of the internal space of the openable container, and the support surface of the target object's bounding box can be accommodated within the rectangle of the openable container's support surface. If this exists, the support surface of the target object is determined to be accommodated in the container's support surface; if it is not accommodated in the current direction, the container's grid surface is rotated by a certain angle, and the above detection process is repeated until the cumulative rotation angle reaches a preset threshold; if there is still no accommodating solution, the target object is determined not to be accommodated in the container.
[0052] The target object and the bottom geometry of the container are discretized into a mesh using a meshing method, and the containment capacity between them is analyzed based on the mesh representation. The implementation steps are as follows:
[0053] Step 5.1.1: Construct a two-dimensional plane using the container bottom surface information, set the mesh size, and convert the plane into a mesh surface;
[0054] Step 5.1.2: Traverse all grids, set the grid value occupied by the container's carrying surface to 1, otherwise set it to 0, to determine the grid area that can be used to place the target item;
[0055] Step 5.1.3: Convert the support surface dimensions of the target object to a grid number dimension using the grid size; detect the number of consecutive grids along the x and y axes of the container's bearing surface area. If the number of consecutive grids at a certain location is greater than the grid number dimension of the target object, then the support surface of the target object can be accommodated on the bottom surface of the container. Figure 2 As shown;
[0056] Step 5.1.4: If the target object's support surface cannot be accommodated on the container's bottom surface in this analysis, rotate the container's mesh surface by a certain angle and repeat steps 5.1.2 and 5.1.3. When the cumulative rotation angle reaches 90° and there is still no accommodating solution, the target object cannot be accommodated in the container. An accommodability test diagram is shown below. Figure 3 As shown.
[0057] Step 5.2: Capacity analysis based on height constraints;
[0058] For the occlusion space formed by a visible object, based on the principle of perspective projection, the occlusion space formed by the visible object on the supporting surface is calculated; it is determined whether the height of the target object at a certain point in the occlusion space is lower than the height of the occlusion space at that position; for areas that meet the height limit, the accommodability analysis based on the supporting surface described in step 5.1 is performed to obtain the possible positions of the target object in the occlusion space. The occlusion space of a visible object is similar to the shadow volume of an object under a point light source. Assuming the position of the viewpoint in the scene is l, the equation of the projected plane is defined as... Let p be the coordinates of a point on a visible object. The relationship between point p and its projection p' on the projection plane can be expressed as: p' = M·p
[0059] Where matrix M is:
[0060] After obtaining the projection of the visible object onto the projection plane, the corresponding occlusion space can be calculated. For example... Figure 4 As shown, S is the camera position, ΔABC is the visible object in the camera, ΔA'B'C' is the plane region Γ occluded by the object, and the polyhedron ABCC'B'A' is the occlusion space formed by ΔABC.
[0061] Unlike openable containers, the height of a visible object relative to its supporting surface varies at different points within the occluded space. It is necessary to determine whether the height of the target object at a certain point within the occluded space is lower than the height of the occluded space at that location. For example... Figure 5 As shown, first, the shadow cast by the visible object on the supporting surface is generated. Then, the shadow area S is set... shadow The height of the interior point relative to the supporting surface is the height of the target object, resulting in a new point cloud S. shadow If S' shadow The projection of a point in the middle under this viewpoint is still in S. shadow If the point is inside, then the point is in S. shadow The corresponding point may accommodate the target object. A target object with a height of 0.05m is located at... Figure 5 Possible locations in the scenario shown in (a) are as follows Figure 5 As shown in (b).
[0062] By performing a planar accommodability analysis on areas that meet height restrictions, the possible locations of the target object in the scene can be determined.
[0063] Step 5.3: Compatibility analysis combining support surface and height constraints: By comprehensively applying the compatibility analysis methods described in Steps 5.1 and 5.2, the compatibility problem of the target object in a specific container is decomposed into the compatibility problem of the object's height relative to the support surface and the compatibility problem of the target object's support surface in the bottom of the container. By combining the two, the probability and potential location of the target object are determined.
[0064] Step 6: Potential Location Ranking: Based on the results of the accommodability analysis, rank all unoccupied grids. The ranking can be based on factors such as the stability of the target object placed on the grid, the positional relationship of the grid with other known objects, etc.
[0065] Step 7: Reasoning Search: Examine each potential location sequentially according to the sorting results. This can be done by directly observing the location or by using a robotic arm to probe it.
[0066] Step 8: Target object location: The search ends when the target object is found at a potential location.
[0067] This algorithm not only relies on visual information, but also combines the spatial attributes of objects and potential location reasoning, which greatly improves the recognition rate of occluded objects and optimizes the search path.
[0068] Example:
[0069] Scenario Description: A drawer contains several items, including a target object (e.g., a screwdriver) and other obstructions (e.g., pliers, wrenches, measuring tape, etc.). The robot needs to locate the screwdriver using visual observation and reasoning, without knowing its exact location.
[0070] Step 1: Image Acquisition and Preprocessing: The robot uses a camera to acquire RGB images of the inside of the drawer and performs preprocessing, such as image denoising and color correction.
[0071] Step 2: Visual object detection and segmentation: Use image segmentation algorithms (e.g., Mask R-CNN) to segment visible objects (including partially visible objects) in the image and identify their categories (e.g., pliers, wrench, measuring tape, etc.).
[0072] Step 3: Discretized Representation: The bottom plane of the drawer is discretized into a two-dimensional grid. Each grid represents a small region.
[0073] Step 4: Calculate the area occupied by visible objects: Based on the segmentation results, calculate the area occupied by each visible object in the grid. Mark the grid area occupied by the object as "occupied".
[0074] Step 5: Capacity Analysis: Perform a capacity analysis for each unoccupied grid cell.
[0075] Step 5.1: Support Surface Determination: Determine whether there is a support surface below the grid. For example, if the grid is below the bottom of a drawer or the top of another object, then a support surface is considered to exist.
[0076] Step 5.2: Height Limitation: Determine if there is sufficient space above the grid to accommodate the target object. Based on the height of the target object and the current grid height, determine whether the target object can be placed on the grid.
[0077] Step 6: Potential Location Ranking: Based on the results of the accommodability analysis, rank all unoccupied grids. The ranking can be based on factors such as the stability of the target object placed on the grid, the positional relationship of the grid with other known objects, etc.
[0078] Step 7: Reasoning Search: Examine each potential location sequentially according to the sorting results. This can be done by directly observing the location or by using a robotic arm to probe it.
[0079] Step 8: Target object location: The search ends when the target object is found at a potential location.
[0080] It is understood that the present invention has been described through some embodiments, and those skilled in the art will recognize that various changes or equivalent substitutions can be made to these features and embodiments without departing from the spirit and scope of the invention. Furthermore, under the teachings of the present invention, these features and embodiments can be modified to adapt to specific situations and materials without departing from the spirit and scope of the invention. Therefore, the present invention is not limited to the specific embodiments disclosed herein, and all embodiments falling within the scope of the claims of this application are within the protection scope of the present invention.
Claims
1. An occluded object reasoning search method based on a grid-discrete representation accommodability analysis method, characterized by, Includes the following steps: Step 1: Image Acquisition and Preprocessing: The robot uses a camera to acquire RGB images of the inside of the container that may hold the object, and then performs preprocessing. Step 2: Visual object detection and segmentation: Use image segmentation algorithms to segment visible objects in the image and identify their categories; Step 3: Mesh Discretization: When an object is placed, the contact surface between the object and other objects is fixed, i.e., the supporting surface; when an object is placed inside a container, the surface supporting the object is usually fixed, i.e., the bearing surface; the bearing surface of the openable container and the supporting surface of the target object are projected onto a two-dimensional plane and discretized using a grid of set size, with each grid representing a small region. Step 4: Calculate the area occupied by visible objects: Based on the segmentation results, calculate the area occupied by each visible object in the grid, mark the grid occupied by the container's carrying surface as occupied, and use the unoccupied grid as candidate positions; Convert the dimensions of the target object's supporting surface to the grid number dimension; Step 5: Capacity Analysis: Perform a capacity analysis for each unoccupied grid. Step 5.1: Compatibility analysis based on the support surface: For openable containers, based on the meshed discrete representation, the number of consecutive unoccupied grids in the X and Y axes of the container's bearing surface is detected to determine whether there is a region where the number of consecutive unoccupied grids is greater than or equal to the size of the target object's grid. A two-dimensional rectangle is used to describe the bearing surface of the openable container. The condition for the target object to be accommodated in the openable container is that the height of the target object is lower than the height of the internal space of the openable container, and the supporting surface of the target object's bounding box can be accommodated within the rectangle of the openable container's bearing surface. If such a region exists, it is determined that the supporting surface of the target object can be accommodated within the container's bearing surface. If it cannot be accommodated in the current direction, rotate the container grid surface by a certain angle and repeat the above detection process until the cumulative rotation angle reaches the preset threshold. If there is still no solution that can accommodate the object, then the object is determined to be unacceptable to be contained in the container. Step 5.2: Capacity analysis based on height constraints; Assume the position of the viewpoint in the scene is The equation of the projected plane is defined as The coordinates of a point on the visible object are The relationship between the point and its projection on the projected plane can be expressed as: where the matrix is: After obtaining the projection of the visible object onto the projection plane, the corresponding occlusion space is calculated. To determine whether a target object's height is lower than the height of the occlusion space at a certain location within the occlusion space: First, generate the shadow cast by the visible object on the supporting surface. Then, set the shadow area... The height of the interior point relative to the supporting surface is the height of the target object, resulting in a new point cloud. ,like The projection of a point in the middle under this viewpoint is still If the point is inside, then the point is within. The corresponding point may accommodate the target item; for areas that meet the height restrictions, try in-plane accommodability analysis to obtain the possible locations of the target item in the scene; Step 5.3: Compatibility analysis combining support surface and height constraints; Step 6: Potential Location Ranking: Based on the results of the accommodation analysis, sort all unoccupied grids; Step 7: Reasoning Search: Examine each potential location sequentially according to the sorting results; Step 8: Target object location: The search ends when the target object is found at a potential location.
2. The occluded object reasoning search method based on the mesh-discrete representation accommodatability analysis method according to claim 1, characterized in that, Step 2 includes extracting 3D point cloud information of visible objects by combining depth images. After obtaining the point cloud of visible objects in the scene, bounding box fitting can be used to obtain the approximate structure, position and pose information of the visible objects.
3. The occluded object reasoning search method based on the mesh-discrete representation accommodatability analysis method according to claim 1, characterized in that, The capacity analysis in step 5 filters candidate containers in two stages: first, selecting openable containers that may store the target object; second, after opening the openable container, determining which visible objects might be occupying the target object. The target object to be searched is represented by OBB. The capacity analysis problem is simplified to whether a cuboid of a given size can be placed inside a polyhedron. The containers in the openable container selection stage are represented by cuboids. In the visible object selection phase, the container is represented by a truncated cone. When an object is placed, the contact surface between the object and other objects is fixed, i.e., the supporting surface. When an object is placed inside the container, the surface supporting the placed object is usually fixed, i.e., the supporting surface. Therefore, the problem of the accommodability of the target object in a specific container is transformed into the problem of the accommodability of the object relative to the height of the supporting surface, and the problem of the accommodability of the target object's supporting surface in the bottom of the container.
4. The occluded object reasoning search method based on the mesh-discrete representation accommodatability analysis method according to claim 1, characterized in that, Step 5.1 uses a meshing method to discretize the bottom geometry of the target object and the container into a mesh, and analyzes the containment capacity between them based on the mesh representation. The steps are as follows: Step 5.1.1: Construct a two-dimensional plane using the container bottom surface information, set the mesh size, and convert the plane into a mesh surface; Step 5.1.2: Traverse all grids, set the grid value occupied by the container's carrying surface to 1, otherwise set it to 0, to determine the grid area that can be used to place the target item; Step 5.1.3: converting the support surface size of the target object into grid number size by using the grid size; detecting the container support surface area axis and the continuous grid number in the axis direction, if the continuous grid number of a certain position is greater than the grid number size of the target object, the support surface of the target object can be accommodated in the container bottom surface; Step 5.1.4: If the target object's support surface cannot be accommodated on the bottom surface of the container in this analysis, rotate the container's grid surface by a certain angle and repeat steps 5.1.2 and 5.1.
3. When the cumulative rotation angle reaches 90° and there is still no solution that can accommodate the target object, the target object cannot be accommodated in the container.
5. The occluded object reasoning search method based on the mesh-discrete representation accommodatability analysis method according to claim 4, characterized in that, Step 5.2, for the occlusion space formed by the visible object, calculates the occlusion space formed by the visible object on the supporting surface based on the principle of perspective projection; determines whether the height of the target object is lower than the height of the occlusion space at that position when it is located at a certain point in the occlusion space; for areas that meet the height limit, performs the same accommodability analysis based on the supporting surface as described in step 5.1 to obtain the possible locations of the target object in the occlusion space.
6. The inferential search method for occluded objects based on the grid discrete representation capacity analysis method according to claim 4, characterized in that, Step 5.3 combines the capacity analysis methods described in steps 5.1 and 5.2 to decompose the capacity problem of the target object in a specific container into the capacity problem of the object's height relative to the bearing surface and the capacity problem of the target object's supporting surface in the bottom of the container. By combining the two, the probability and potential location of the target object are determined.
7. The occluded object reasoning search method based on the mesh-discrete representation accommodatability analysis method according to claim 1, characterized in that, The sorting in step 6 is based on the stability of the target object placed on the grid and the positional relationship between the grid and other known objects.