An acceptance method and device based on 3D point cloud data labeling and a storage medium

Abstract

The present application belongs to the technical field of data labeling, and particularly relates to a 3D point cloud data labeling-based acceptance method and device and a storage medium. The method comprises the following steps: S1, standard presetting, in which, based on an acceptance object and an acceptance standard, a size of an element box and a fitting error of the element box are preset in advance; S2, system judgment, in which, according to a comparison between the standard presetting and actual labeling, it is judged whether the size of the element box is correct, whether the fitting of the element box is within the error range, whether the element boxes overlap, and an abnormal position and an abnormal reason are identified; and S3, inspection, in which, according to the abnormal position and the abnormal reason judged by the system, a form that can be viewed visually is used to display, and the abnormal position and the abnormal reason are clicked to inspect. The purpose is to solve the problem of misjudgment and low efficiency in the background art.

Description

Technical Field

[0001] This invention belongs to the field of data annotation technology, specifically relating to an acceptance method, device and storage medium based on 3D point cloud data annotation. Background Technology

[0002] The field of 3D point cloud data annotation is booming, and data acceptance is a crucial part of the process. Conventional data acceptance methods typically involve manual verification by personnel, checking aspects such as element bounding box size, overlap between boxes, and box alignment. However, a single frame of data can contain dozens of elements, making human verification inevitably prone to misjudgments. Furthermore, manual verification of element bounding boxes is inefficient, limiting the output of the acceptance process. Summary of the Invention

[0003] The purpose of this invention is to provide an acceptance method, device, and storage medium based on 3D point cloud data annotation, in order to solve the problems of misjudgment and low efficiency of manual verification as pointed out in the background art.

[0004] To achieve the above-mentioned technical objectives, the technical solution adopted by the present invention is as follows:

[0005] Firstly, this application discloses an acceptance method based on 3D point cloud data annotation, comprising the following steps:

[0006] S1, Standard Preset: In the standard preset, based on the acceptance object and acceptance standard, the size of the element box and the fitting error of the element box are preset.

[0007] S2. The system judges whether the size of the element box is in line with the standard preset and the actual annotation, whether the fitting of the element box is within the error range, whether the element boxes overlap, and marks the abnormality and the cause of the abnormality.

[0008] S3. Check: Based on the system's judgment of the anomaly and its cause, display it in a visually viewable form, and click to check the anomaly and its cause.

[0009] In conjunction with the first aspect, in some alternative implementations, the method further includes,

[0010] When setting the size of the element box, the length, width, and height of the element box can be customized according to different acceptance objects and acceptance standards.

[0011] In conjunction with the first aspect, in some alternative implementations, the method further includes,

[0012] When determining the fitting error range of an element box, it iterates through all points in the element box and checks whether the shortest distance from the box's face is within the preset error range. If it is outside the error range, it is marked as an abnormality.

[0013] In conjunction with the first aspect, in some alternative implementations, the method further includes,

[0014] The preset error value is w. In a three-dimensional coordinate system, let S be a face of the element box, represented as: Ax + By + Cz + D = 0; vector n = (A, B, C) is the normal vector of plane S. Let the coordinates of a point P be (x1, y1, z1), and the distance from point P to a face be... Iterate through all points in the element box. If dmin ≤ w, it is determined that the current element box meets the fitting requirements.

[0015] In conjunction with the first aspect, in some alternative implementations, the method further includes,

[0016] When determining whether two element boxes overlap, the system first checks if a point is inside an element box, and then checks if that point exists in another element box. If it does not exist, the system does not respond. If the point exists in another element box, it indicates that the element boxes overlap.

[0017] In conjunction with the first aspect, in some alternative implementations, the method further includes,

[0018] When determining whether a point is within the element box, the angle relationship between the point and the vector formed by the eight vertices of the 3D solid space where the point is located and the vectors of each edge is determined. The inner product operation is performed based on the angle relationship. If the inner product is negative, the point is determined to be outside the element box.

[0019] In conjunction with the first aspect, in some alternative implementations, the method further includes,

[0020] In the inner product operation, when the vertices are A and B and the point is P, we set vector AP = (x1, y1) and vector AB = (x2, y2). Based on the angle between vectors AP and AB, the inner product operation is expressed as: AP * AB = x1x2 + y1y2.

[0021] Secondly, this application also discloses an acceptance device based on 3D point cloud data annotation, the device comprising:

[0022] The preset module is used to preset, but is not limited to, the size of the element box and the fitting error of the element box;

[0023] The judgment module is used to compare the data in the preset module with the actual annotation to determine whether the size of the element box is correct, whether the alignment of the element box is within the error range, whether the element boxes overlap, and to mark the abnormality and the cause of the abnormality.

[0024] The inspection module is used to inspect and determine the anomalies and their causes as identified by the judgment module.

[0025] In conjunction with the second aspect, in some alternative embodiments, the apparatus further includes:

[0026] The standard module is used to store different acceptance objects and their acceptance criteria.

[0027] Thirdly, this application also discloses a computer-readable storage medium storing a computer program that, when run on a computer, causes the computer to perform the method described above.

[0028] The invention employing the above technical solution has the following advantages:

[0029] Based on different acceptance objects and acceptance standards, the size and fitting error of the element boxes for this annotation are preset. The preset standards are based on the acceptance standards, and the system then judges whether the actual annotation matches the preset standards. If they match, the acceptance is passed; if they do not match, they are marked as abnormal, thus reducing misjudgments. The system can also judge whether the element boxes overlap, which can further reduce misjudgments. Abnormalities are then judged by the acceptance personnel, which improves the acceptance efficiency and reduces the workload of the acceptance personnel. Attached Figure Description

[0030] This application can be further illustrated by the non-limiting embodiments given in the accompanying drawings. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as limiting the scope. For those skilled in the art, other related drawings can be obtained from these drawings without creative effort.

[0031] Figure 1 Flowchart of the acceptance method provided in the embodiments of this application Figure 1 ;

[0032] Figure 2 A schematic diagram illustrating whether the element boxes overlap in the acceptance method provided in the embodiments of this application;

[0033] Figure 3 Flowchart of the acceptance method provided in the embodiments of this application Figure 2 ;

[0034] Figure 4A block diagram of the acceptance device provided in the embodiments of this application;

[0035] The symbols for the main components are explained below:

[0036] Acceptance device 200, preset module 210, judgment module 220, inspection module 230. Detailed Implementation

[0037] The present application will be described in detail below with reference to the accompanying drawings and specific embodiments. It should be noted that similar or identical parts are referred to by the same reference numerals in the drawings or description. Implementations not shown or described in the drawings are forms known to those skilled in the art. In the description of this application, terms such as "first" and "second" are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0038] Please refer to the attached document. Figure 1 This application provides an acceptance method based on 3D point cloud data annotation, including the following steps:

[0039] Step 110, Standard Preset: In the standard preset, based on the acceptance object and acceptance criteria, presets include but are not limited to the size of the element frame and the fitting error of the element frame.

[0040] Step 120: The system judges whether the size of the element box is in line with the standard preset and the actual annotation, whether the fitting of the element box is within the error range, whether the element boxes overlap, and marks the abnormality and the cause of the abnormality.

[0041] Step 130: Check. Based on the anomalies and their causes identified by the system, display them in a visually readable format, and click to check the anomalies and their causes.

[0042] In the above implementation, based on different acceptance objects and acceptance standards, preset standards are set. By setting the standard in advance, the annotator can mark according to the standard, which can improve the annotator's marking efficiency and reduce the possibility of misjudgment. Furthermore, by comparing the annotator's actual marking with the preset standard, the possibility of misjudgment can be further reduced, and the annotator's marking abnormalities and their causes can be identified. Then, by having the verification personnel check the anomalies and their causes determined by the system, the work efficiency of the verification personnel can be effectively improved, and the output of acceptance can be increased.

[0043] It's understandable that 3D point cloud data annotation, when scanned by scanning instruments such as LiDAR, may not include all the points of the stereoscopic image of the object to be annotated. Sometimes, the scan may only show one face of the object, which could lead to misjudgment by the annotator. For example, when a scanning instrument scans a truck from behind, it may only show four points on the rear of the truck, i.e., the width and height, but not the full length. This could cause the annotator to misidentify the truck. By setting annotation standards in advance, the possibility of misjudgment can be reduced, thereby improving annotation efficiency.

[0044] As an optional implementation, the method may further include,

[0045] In step 110, when setting the size of the element box, the length, width, and height of the element box are customized according to different acceptance objects and acceptance standards.

[0046] In this embodiment, the acceptance objects can be vehicles, road information, static objects, and dynamic objects. It is understood that vehicles include, but are not limited to, cars, vans, trucks, bicycles, etc.; road information includes, but is not limited to, road length, width, sidewalks, signs, etc.; static objects include, but are not limited to, green belts, trees, guardrails, streetlights, billboards, etc.; and dynamic objects include, but are not limited to, pedestrians, cats, dogs, etc. Pre-setting the element box size of the acceptance object facilitates the identification and annotation by the annotator.

[0047] As an optional implementation, the method may further include,

[0048] In step 120, when determining the fitting error range of the element box, all points in the element box are traversed, and it is determined whether the shortest distance from the surface of the box is within the preset error range. If it is outside the error range, it is marked as an abnormality.

[0049] Understandably, by iterating through the points marked by the annotator within the element box and comparing them with the shortest distance of the preset element box face, if the shortest distance is within the preset error range, the element box marked by the annotator is judged to meet the acceptance standard; if the shortest distance exceeds the preset error range, it is judged to be an abnormality and marked.

[0050] In this embodiment, the preset error value is w, which can be set according to different acceptance objects and acceptance standards. When specifically judging whether the element box marked by the annotator is within the error range, the following method can be used: In a three-dimensional coordinate system, let a certain face of the element box be S, represented as: Ax + By + Cz + D = 0; vector n = (A, B, C) is the normal vector of plane S, and let the coordinates of a certain point P be (x1, y1, z1), and the distance from point P to a certain face be... Iterate through all points in the element box. If dmin ≤ w, then the current element box meets the fitting requirements.

[0051] As an optional implementation, the method may further include,

[0052] In step 120, when determining whether element boxes overlap, it is determined whether a certain point is inside an element box, and then it is determined whether the point exists in other element boxes. If it does not exist, the system does not respond; if the point exists in other element boxes, it indicates that there is an overlap between element boxes.

[0053] It is understandable that when a point within an element box exists both within the current element box and within another element box, it means that the two element boxes at least partially overlap.

[0054] Please refer to the attached document. Figure 2 In this embodiment, when determining whether a point is within the element box, the angle relationship between the vector formed by the point and the eight vertices of the 3D solid space where the point is located and the vector of each edge is determined. The inner product operation is performed based on the angle relationship. If the inner product is negative, the point is determined to be outside the element box.

[0055] Understandably, in the inner product operation, let's take vertices A and B, and point P as an example. Let's set vector AP = (x1, y1) and vector AB = (x2, y2). Based on the angle between vectors AP and AB, the inner product operation is expressed as: AP * AB = x1x2 + y1y2. If the inner product is negative, it's determined that the point is outside the element box, and the process ends. If the inner product is positive, the inner product calculation continues for the next point. This process iterates through all points. When the inner product of all points is positive, it's determined that all points are inside the element box and do not exist within other element boxes, meaning there is no overlap between element boxes.

[0056] Please refer to the appendix. Figure 2-3 The acceptance method based on 3D point cloud data annotation is described in detail below:

[0057] S1. The acceptance personnel shall, based on the objects and acceptance standards to be accepted, pre-set the length, height and width dimensions of the element box according to the acceptance standards;

[0058] S2. Based on the objects and standards to be inspected, the inspection personnel shall pre-set the fitting error range w between the element boxes.

[0059] S3. The system compares the length, height, and width of the element box actually annotated by the annotator with the pre-set length, height, and width of the element box. If there are any abnormalities, the system marks the abnormality and the reason for the abnormality.

[0060] S4. The system compares the actual fitting distance of the element box marked by the annotator within the element box with the pre-set fitting error range w. The specific comparison is as follows: In the three-dimensional coordinate system, let a certain face of the element box be S, represented as: Ax + By + Cz + D = 0; vector n = (A, B, C) is the normal vector of plane S, and let the coordinates of a certain point P be (x1, y1, z1), and the distance from point P to a certain face be... Iterate through all points in the element box. If dmin ≤ w, it is determined that the current element box meets the fitting requirements, that is, the actual fitting distance is within the preset error range w. If dmin > w, it is determined that the actual fitting distance exceeds the preset error range w, and there is a fitting abnormality between the two element boxes. The system marks this as an abnormality and indicates the reason for the abnormality.

[0061] S5. The system verifies whether there is overlap between element boxes. The specific steps are as follows: First, determine whether a point is inside an element box. Then, determine whether the point is inside other element boxes. When determining whether a point is inside an element box, determine the angle relationship between the vector formed by the point and the eight vertices of the 3D solid space where the point is located, and the vectors of each edge. Perform an inner product operation based on the angle relationship. Taking vertices A and B and the point P as an example, set vector AP = (x1, y1) and vector AB = (x2, y2). Based on the angle relationship between vectors AP and AB, the inner product operation is expressed as: AP * AB = x1x2 + y1y2. If the inner product is negative, it is determined that the point is outside the element box; if the inner product is positive, the inner product calculation continues for the next point. This process is repeated for all points. When the inner product of all points is positive, it is determined that all points are inside the element boxes and do not exist in other element boxes, and there is no overlap between element boxes.

[0062] S6. The acceptance personnel will view the anomalies and their causes as determined by the system in a visually readable format and click to verify the anomalies and their causes.

[0063] Based on the above design, the size and fitting error of the element box during the annotation can be preset according to different acceptance objects and acceptance standards. The preset standard is based on the acceptance standard, and then the system judges whether the actual annotation is consistent with the preset standard. If it is consistent, the acceptance is passed; if it is inconsistent, it is marked as an anomaly, thereby reducing misjudgment. In addition, the system can also judge whether the element boxes overlap, which can further reduce misjudgment. The anomalies are then judged by the acceptance personnel, which improves the acceptance efficiency and reduces the workload of the acceptance personnel.

[0064] Please refer to the attached document. Figure 4This application also provides an acceptance device based on 3D point cloud data annotation. The acceptance device 200 includes at least one software functional module that can be stored in a storage module or embedded in an operating system (OS) in the form of software or firmware. For example, the software functional modules and computer programs included in the acceptance device 200.

[0065] The acceptance device 200 may include a preset module 210, a judgment module 220, and an inspection module 230. The functions of each unit may be as follows:

[0066] The preset module 210 is used to preset, but is not limited to, the size of the element box and the fitting error of the element box;

[0067] The judgment module 220 is used to compare the data in the preset module with the actual annotation to determine whether the size of the element box is in line with the actual annotation, whether the fitting of the element box is within the error range, whether the element boxes overlap, and to mark the abnormality and the cause of the abnormality.

[0068] The inspection module 230 is used to inspect the anomalies and causes identified by the judgment module.

[0069] Optionally, the acceptance device 200 also includes a standard module, which stores different acceptance objects and acceptance standards to provide reference standards for the pre-set size of the element frame and the fitting error of the element frame.

[0070] Optionally, the inspection module 230 can also be used to display the anomalies and their causes in a list format.

[0071] In this embodiment, the storage module can be, but is not limited to, random access memory, read-only memory, programmable read-only memory, erasable programmable read-only memory, electrically erasable programmable read-only memory, etc. In this embodiment, the storage module can be used to store the working status of the judgment module 220 and the verification module 230. Of course, the storage module can also be used to store programs, which the processing module executes after receiving an execution instruction.

[0072] This application also provides a computer-readable storage medium. The computer-readable storage medium stores a computer program that, when run on a computer, causes the computer to perform the acceptance method described in the above embodiments.

[0073] Based on the above description of the embodiments, those skilled in the art can clearly understand that this application can be implemented by hardware or by using software plus necessary general-purpose hardware platforms. Based on this understanding, the technical solution of this application can be embodied in the form of a software product. The software product can be stored in a non-volatile storage medium (such as CD-ROM, USB flash drive, mobile hard drive, etc.) and includes several instructions to cause a computer device (such as a personal computer, braking device, or network device, etc.) to execute the methods described in the various implementation scenarios of this application.

[0074] In summary, this application provides an acceptance method, apparatus, and storage medium based on 3D point cloud data annotation. In this solution, the size and fitting error of the element bounding boxes during annotation are pre-set according to different acceptance objects and standards. The pre-set standards are based on the acceptance criteria. The system then judges whether the actual annotation conforms to the pre-set standards. If they conform, the acceptance is passed; otherwise, it is marked as an anomaly, thereby reducing misjudgments. Furthermore, the system can also determine whether element bounding boxes overlap, further reducing misjudgments. Anomalies are then judged by acceptance personnel, improving acceptance efficiency and reducing the workload of acceptance personnel.

[0075] In the embodiments provided in this application, it should be understood that the disclosed apparatus, systems, and methods can also be implemented in other ways. The apparatus, systems, and methods embodiments described above are merely illustrative. For example, the flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code, which includes one or more executable instructions for implementing a specified logical function. It should also be noted that each block in a block diagram and / or flowchart, and combinations of blocks in block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions. Furthermore, the functional modules in the various embodiments of this application can be integrated together to form an independent part, or each module can exist independently, or two or more modules can be integrated to form an independent part.

[0076] The above description is merely an embodiment of this application and is not intended to limit the scope of protection of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. An acceptance method based on 3D point cloud data annotation, characterized in that: Includes the following steps: S1, Standard Preset: In the standard preset, based on the acceptance object and acceptance standard, the size of the element box and the fitting error of the element box are preset. S2. The system judges whether the size of the element box is in line with the standard preset and the actual annotation, whether the fitting of the element box is within the error range, whether the element boxes overlap, and marks the abnormality and the cause of the abnormality. When determining the fitting error range of an element box, iterate through all points in the element box and determine whether the shortest distance from the box's face is within the preset error range. If it is outside the error range, mark it as an abnormal point. The preset error value is w. In a three-dimensional coordinate system, let S be a face of the element box, represented as: Ax + By + Cz + D = 0; vector n = (A, B, C) is the normal vector of plane S. Let the coordinates of a point P be (x1, y1, z1), and the distance from point P to a face be... Iterate through all points in the element box, and if dmin ≤ w, determine that the current element box meets the fitting requirements; S3. Check: Based on the system's judgment of the anomalies and their causes, display them in a visually readable format, and click to check the anomalies and their causes.

2. The acceptance method based on 3D point cloud data annotation according to claim 1, characterized in that: The method also includes, When setting the size of the element box, the length, width, and height of the element box can be customized according to different acceptance objects and acceptance standards.

3. The acceptance method based on 3D point cloud data annotation according to claim 1, characterized in that: The method also includes, When determining whether two element boxes overlap, the system first checks if a point is inside an element box, and then checks if that point exists in another element box. If it does not exist, the system does not respond. If the point exists in another element box, it indicates that the element boxes overlap.

4. The acceptance method based on 3D point cloud data annotation according to claim 3, characterized in that: The method also includes, When determining whether a point is within the element box, the angle relationship between the point and the vector formed by the eight vertices of the 3D solid space where the point is located and the vectors of each edge is determined. The inner product operation is performed based on the angle relationship. If the inner product is negative, the point is determined to be outside the element box.

5. The acceptance method based on 3D point cloud data annotation according to claim 4, characterized in that: The method also includes, In the inner product operation, when the vertices are A and B and the point is P, we set vector AP = (x1, y1) and vector AB = (x2, y2). Based on the angle between vectors AP and AB, the inner product operation is expressed as: AP * AB = x1x2 + y1y2.

6. An acceptance device based on 3D point cloud data annotation, used to execute the acceptance method based on 3D point cloud data annotation as described in any one of claims 1-5, characterized in that, The device includes: The preset module is used to preset, but is not limited to, the size of the element box and the fitting error of the element box; The judgment module is used to compare the data in the preset module with the actual annotation to determine whether the size of the element box is correct, whether the alignment of the element box is within the error range, whether the element boxes overlap, and to mark the abnormality and the cause of the abnormality. When determining the fitting error range of the element frame, all points within the element frame are traversed, and it is determined whether the shortest distance from the frame's face is within a preset error range. If it is outside the error range, it is marked as an anomaly. The preset error value is w. In a three-dimensional coordinate system, let a face of the element frame be S, represented as: Ax + By + Cz + D = 0; vector n = (A, B, C) is the normal vector of plane S, and let the coordinates of a point P be (x1, y1, z1), and the distance from point P to a face be... Iterate through all points in the element box, and if dmin ≤ w, determine that the current element box meets the fitting requirements; The inspection module is used to inspect and determine the anomalies and their causes as identified by the judgment module.

7. The acceptance device based on 3D point cloud data annotation according to claim 6, characterized in that: The device further includes: The standard module is used to store different acceptance objects and their acceptance criteria.

8. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when run on a computer, causes the computer to perform the method as described in any one of claims 1-5.

Images