Data processing method and device, electronic equipment and computer readable medium

Abstract

Embodiments of the present disclosure disclose a data processing method, device, electronic equipment and medium. A specific implementation of the method comprises: acquiring a point cloud data sequence used for constructing a point cloud map; generating an object recognition result list corresponding to the point cloud data sequence; performing deletion processing on target object information in the object recognition result list and each recognition result corresponding to the target object information to obtain a processed object recognition result list; and performing deletion processing on a target point cloud set in each frame of point cloud data in the point cloud data sequence according to the processed object recognition result list to generate deleted point cloud data, thereby obtaining a deleted point cloud data sequence, wherein the target point cloud set represents a moving obstacle object in the each frame of point cloud data. The implementation can quickly and efficiently delete the target point cloud set associated with the moving obstacle object in each frame of point cloud data in the point cloud data sequence.

Description

Technical Field

[0001] Embodiments of this disclosure relate to the field of computer technology, and more particularly to data processing methods, apparatus, electronic devices, and computer-readable media. Background Technology

[0002] Currently, autonomous driving products often rely on high-precision point cloud maps. When collecting data for these high-precision point cloud maps, noise from people, vehicles, and other objects on the road is introduced. These moving, dynamic obstacles negatively impact the use of high-precision point cloud maps and need to be removed. A common approach is to use deep learning models to identify moving obstacles and then remove the associated point cloud sets.

[0003] However, when using the above method to remove point clouds, the following technical problems often arise:

[0004] When using deep learning models to identify moving obstacles, the accuracy and recall rates often fall short of 100%, potentially resulting in the deletion or retention of movable obstacles. This is especially true for moving vehicles, where even a few frames left behind can produce noticeable motion blur, making noise filtering ineffective. Summary of the Invention

[0005] The summary portion of this disclosure is intended to provide a brief overview of the concepts, which will be described in detail in the detailed description portion. This summary portion is not intended to identify key or essential features of the claimed technical solutions, nor is it intended to limit the scope of the claimed technical solutions.

[0006] Some embodiments of this disclosure provide data processing methods, apparatuses, electronic devices, and computer-readable media to address one of the technical problems mentioned in the background section above.

[0007] In a first aspect, some embodiments of this disclosure provide a data processing method, including: acquiring a point cloud data sequence for constructing a point cloud map; generating an object recognition result list corresponding to the point cloud data sequence, wherein the object recognition result list includes information on each identified object and recognition results for each object information; deleting target object information and recognition results corresponding to the target object information from the object recognition result list to obtain a processed object recognition result list; and deleting target point clouds in each frame of point cloud data in the point cloud data sequence according to the processed object recognition result list to generate deleted point cloud data, thereby obtaining a deleted point cloud data sequence, wherein the target point clouds represent moving obstacle objects in each frame of point cloud data.

[0008] Optionally, generating the object recognition result list corresponding to the point cloud data sequence includes: in response to the target point cloud data in the point cloud data sequence not being the first frame of point cloud data in the point cloud data sequence, performing the following object recognition result list generation steps: determining at least one recognition result of the target point cloud data; integrating the at least one recognition result into an initial object recognition result list according to preset conditions to obtain an integrated initial object recognition result list, wherein the initial object recognition result list is related to the point cloud data sequence preceding the target point cloud data in the point cloud data sequence; determining the integrated initial object recognition result list as the initial object recognition result list; determining whether the target point cloud data has a next frame of point cloud data; and in response to the absence of a next frame of point cloud data, determining the initial object recognition result list as the object recognition result list.

[0009] Optionally, the above method further includes: in response to the existence, determining the next frame of point cloud data of the target point cloud data as the target point cloud data, and continuing to execute the above object recognition result list generation step.

[0010] Optionally, the above method further includes: in response to the target point cloud data being the first frame of point cloud data in the point cloud data sequence, writing at least one object recognition result corresponding to the first frame of point cloud data into a pre-established empty initial object recognition result list and determining the next frame of point cloud data of the first frame of point cloud data as the target point cloud data, and continuing to execute the object recognition result list generation step.

[0011] Optionally, the above recognition results include: the number of times the object information is recognized and the recognition scores of each object information; and the above-mentioned deletion process of the target object information and the recognition results corresponding to the target object information in the above object recognition result list to obtain the processed object recognition result list includes: determining the number of times each object information is recognized and the highest recognition score among the corresponding recognition scores in each object information according to the above object recognition result list; removing the target object information whose number of recognitions is less than a first threshold and / or whose highest recognition score is less than a second threshold and the recognition results corresponding to the target object information from the above object recognition result list to obtain the processed object recognition result list.

[0012] Optionally, the above-mentioned method of deleting target point clouds in each frame of point cloud data in the above-mentioned point cloud data sequence to generate deleted point cloud data, based on the above-mentioned processed object recognition result list, includes: determining the occurrence time of each object information according to the above-mentioned processed object recognition result list; performing the following operations on each frame of point cloud data according to the forward order of the above-mentioned point cloud data sequence: determining the time difference between the time corresponding to each frame of point cloud data and the occurrence time of each object information in the above-mentioned object information; superimposing the bounding box corresponding to each object information onto the corresponding position in each frame of point cloud data according to the above-mentioned time difference; determining whether there is a point cloud in at least one bounding box in each frame of point cloud data; in response to determining that there is, adding the geometric center information of the bounding box associated with the above-mentioned object information in the corresponding point cloud data to each recognition result of each object information in the above-mentioned object recognition result list, to obtain the added object recognition result list.

[0013] Optionally, the above method further includes: performing the following operations on each frame of point cloud data according to the reverse order of the point cloud data sequence: determining the time difference between the time corresponding to each frame of point cloud data and the time of occurrence of each object information in each object information; superimposing the bounding box corresponding to each object information onto the corresponding position in each frame of point cloud data according to the time difference; determining whether there is a point cloud in at least one bounding box in each frame of point cloud data; in response to determining that there is, adding the geometric center information of the bounding box associated with the object information in the corresponding point cloud data to each recognition result of each object information in the added object recognition result list.

[0014] Optionally, the above method further includes: deleting the target point cloud set in each frame of point cloud data in the above point cloud data sequence based on the geometric center information of the bounding box corresponding to each frame of point cloud data and the volume information of the bounding box to generate deleted point cloud data, thereby obtaining the deleted point cloud data sequence.

[0015] Optionally, the above method further includes: storing the deleted point cloud data sequence for subsequent construction of the point cloud map.

[0016] Secondly, some embodiments of this disclosure provide a data processing apparatus, including: an acquisition unit configured to acquire a point cloud data sequence for constructing a point cloud map; a generation unit configured to generate an object recognition result list corresponding to the point cloud data sequence, wherein the object recognition result list includes information on each identified object and a recognition result for each object information; a first processing unit configured to delete target object information and the recognition results corresponding to the target object information in the object recognition result list to obtain a processed object recognition result list; and a second processing unit configured to delete target point cloud sets in each frame of point cloud data in the point cloud data sequence according to the processed object recognition result list to generate deleted point cloud data, thereby obtaining a deleted point cloud data sequence, wherein the target point cloud sets represent moving obstacle objects in each frame of point cloud data.

[0017] Optionally, the generation unit is further configured to: in response to the target point cloud data in the point cloud data sequence not being the first frame of point cloud data in the point cloud data sequence, perform the following object recognition result list generation steps: determine at least one recognition result of the target point cloud data; integrate the at least one recognition result into an initial object recognition result list according to preset conditions to obtain an integrated initial object recognition result list, wherein the initial object recognition result list is related to the point cloud data sequence preceding the target point cloud data in the point cloud data sequence; determine the integrated initial object recognition result list as the initial object recognition result list; determine whether the target point cloud data has a next frame of point cloud data; and in response to the absence of a next frame of point cloud data, determine the initial object recognition result list as the object recognition result list.

[0018] Optionally, the above apparatus further includes: in response to the presence of the target point cloud data, determining the next frame of point cloud data as the target point cloud data, and continuing to execute the object recognition result list generation step.

[0019] Optionally, the above apparatus further includes: in response to the target point cloud data being the first frame of point cloud data in the point cloud data sequence, writing at least one object recognition result corresponding to the first frame of point cloud data into a pre-established empty initial object recognition result list, and determining the next frame of point cloud data of the first frame of point cloud data as the target point cloud data, and continuing to execute the object recognition result list generation step.

[0020] Optionally, the above recognition results include: the number of times the object information is recognized and the recognition scores of each object information; and the first processing unit is further configured to: determine the number of times each object information is recognized and the highest recognition score among the corresponding recognition scores according to the above object recognition result list; remove the target object information whose number of recognitions is less than a first threshold and / or whose highest recognition score is less than a second threshold and the recognition results corresponding to the target object information from the above object recognition result list to obtain the processed object recognition result list.

[0021] Optionally, the above-mentioned method of deleting target point clouds in each frame of point cloud data in the above-mentioned point cloud data sequence to generate deleted point cloud data, based on the above-mentioned processed object recognition result list, includes: determining the occurrence time of each object information according to the above-mentioned processed object recognition result list; performing the following operations on each frame of point cloud data according to the forward order of the above-mentioned point cloud data sequence: determining the time difference between the time corresponding to each frame of point cloud data and the occurrence time of each object information in the above-mentioned object information; superimposing the bounding box corresponding to each object information onto the corresponding position in each frame of point cloud data according to the above-mentioned time difference; determining whether there is a point cloud in at least one bounding box in each frame of point cloud data; in response to determining that there is, adding the geometric center information of the bounding box associated with the above-mentioned object information in the corresponding point cloud data to each recognition result of each object information in the above-mentioned object recognition result list, to obtain the added object recognition result list.

[0022] Optionally, the second processing unit is further configured to: perform the following operations on each frame of point cloud data according to the reverse order of the point cloud data sequence: determine the time difference between the time corresponding to each frame of point cloud data and the time of occurrence of each object information in each object information; according to the time difference, superimpose the bounding box corresponding to each object information onto the corresponding position in each frame of point cloud data; determine whether there is a point cloud in at least one bounding box in each frame of point cloud data; in response to determining that there is, add the geometric center information of the bounding box associated with the object information in the corresponding point cloud data to each recognition result of each object information in the added object recognition result list.

[0023] Optionally, the second processing unit is further configured to: perform deletion processing on the target point cloud set in each frame of point cloud data in the point cloud data sequence according to the geometric center information of the bounding box corresponding to each frame of point cloud data and the volume information of the bounding box to generate deleted point cloud data, thereby obtaining the deleted point cloud data sequence.

[0024] Optionally, the above-mentioned apparatus further includes: storing the deleted point cloud data sequence for subsequent construction of the point cloud map.

[0025] Thirdly, some embodiments of this disclosure provide an electronic device, including: one or more processors; and a storage device having one or more programs stored thereon, such that when the one or more programs are executed by the one or more processors, the one or more processors implement the method as described in any implementation of the first aspect.

[0026] Fourthly, some embodiments of this disclosure provide a computer-readable medium having a computer program stored thereon, wherein the program, when executed by a processor, implements the method as described in any implementation of the first aspect.

[0027] The above embodiments of this disclosure have the following beneficial effects: The data processing methods of some embodiments of this disclosure can quickly and efficiently delete target point clouds associated with moving obstacle objects in each frame of point cloud data in a point cloud data sequence. Specifically, when identifying moving obstacle objects based on a deep learning model, the accuracy and recall rate of identification often cannot reach 100%, so there may be erroneous deletions or remaining movable obstacle objects. Especially for moving vehicles, even if a few frames are left, obvious ghosting will still occur, and the noise filtering effect is not good. Based on this, the data processing methods of some embodiments of this disclosure can first obtain a point cloud data sequence for constructing a point cloud map. Then, an object recognition result list corresponding to the above point cloud data sequence is generated. The above object recognition result list includes the information of each identified object and the recognition result of each object information. Here, by generating the object recognition result list, the objects associated with the point cloud data sequence can be effectively identified. This can be used for the subsequent determination of moving obstacle objects. Furthermore, the target object information and the recognition results corresponding to the target object information in the above object recognition result list are deleted to obtain the processed object recognition result list. Here, deleting the target object information and its corresponding recognition results from the object recognition result list effectively excludes objects that are not moving obstacles. Finally, based on the processed object recognition result list, the target point cloud set in each frame of the point cloud data sequence is deleted to generate a deleted point cloud data sequence. Here, the target point cloud set represents the moving obstacle object in each frame of the point cloud data. Therefore, this data processing method can quickly and efficiently delete the target point cloud set associated with moving obstacle objects from each frame of the point cloud data sequence. Attached Figure Description

[0028] The above and other features, advantages, and aspects of the embodiments of this disclosure will become more apparent from the accompanying drawings and the following detailed description. Throughout the drawings, the same or similar reference numerals denote the same or similar elements. It should be understood that the drawings are schematic, and the originals and elements are not necessarily drawn to scale.

[0029] Figure 1-2 This is a schematic diagram illustrating an application scenario of a data processing method according to some embodiments of the present disclosure;

[0030] Figure 3 These are flowcharts of some embodiments of the data processing method according to this disclosure;

[0031] Figure 4 These are flowcharts of other embodiments of the data processing methods according to this disclosure;

[0032] Figure 5 These are schematic diagrams illustrating the structure of some embodiments of the data processing apparatus according to this disclosure;

[0033] Figure 6 This is a schematic diagram of the structure of an electronic device suitable for implementing some embodiments of the present disclosure. Detailed Implementation

[0034] Embodiments of this disclosure will now be described in more detail with reference to the accompanying drawings. While some embodiments of this disclosure are shown in the drawings, it should be understood that this disclosure can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided to provide a more thorough and complete understanding of this disclosure. It should be understood that the accompanying drawings and embodiments of this disclosure are for illustrative purposes only and are not intended to limit the scope of protection of this disclosure.

[0035] It should also be noted that, for ease of description, only the parts relevant to the invention are shown in the accompanying drawings. Unless otherwise specified, the embodiments and features described in this disclosure can be combined with each other.

[0036] It should be noted that the concepts of "first" and "second" mentioned in this disclosure are used only to distinguish different devices, modules or units, and are not used to limit the order of functions performed by these devices, modules or units or their interdependencies.

[0037] It should be noted that the terms "a" and "a plurality of" used in this disclosure are illustrative rather than restrictive, and those skilled in the art should understand that, unless otherwise expressly indicated in the context, they should be understood as "one or more".

[0038] The names of messages or information exchanged between multiple devices in the embodiments of this disclosure are for illustrative purposes only and are not intended to limit the scope of such messages or information.

[0039] This disclosure will now be described in detail with reference to the accompanying drawings and embodiments.

[0040] Figures 1-2 This is a schematic diagram illustrating an application scenario of a data processing method according to some embodiments of the present disclosure.

[0041] exist Figures 1-2In this application scenario, electronic device 101 can first acquire point cloud data sequence 103 for constructing point cloud map 102. In this application scenario, the point cloud data sequence 103 may include: first point cloud data 1031, second point cloud data 1032, and third point cloud data 1033. Then, an object recognition result list 104 corresponding to the point cloud data sequence 102 is generated. The object recognition result list 104 includes each identified object and each recognition result for each object information. In this application scenario, the object information may include: pedestrian information, first target vehicle information, and second target vehicle information. In the object recognition result list 104, the recognition result corresponding to pedestrian information may include: first recognition result, second recognition result, and empty. The recognition result corresponding to first target vehicle information may include: third recognition result, fourth recognition result, and empty. The recognition result corresponding to second target vehicle information may include: empty, fifth recognition result, and sixth recognition result. The empty results corresponding to the first, third, and second target vehicle information are the recognition results of each object information associated with the first point cloud data. The second, fourth, and fifth recognition results are the recognition results of various object information associated with the second point cloud data. The blanks corresponding to the pedestrian information, the blanks corresponding to the first target vehicle information, and the sixth recognition result are the recognition results of various object information associated with the third point cloud data. Furthermore, the target object information and the corresponding recognition results in the object recognition result list 104 are deleted to obtain the processed object recognition result list 105. In this application scenario, the target object information can be the second target vehicle information. Therefore, deleting the second target vehicle information and the corresponding recognition results from the object recognition result list 104 yields the processed object recognition result list 105. Finally, based on the processed object recognition result list 105, the target point cloud set in each frame of point cloud data in the point cloud data sequence 103 is deleted to generate deleted point cloud data, resulting in the deleted point cloud data sequence 106. The target point cloud set represents the moving obstacle object in each frame of point cloud data. In this application scenario, the target point cloud set in the first point cloud data of the aforementioned point cloud data sequence 103 is deleted, resulting in the deleted first point cloud data 1061 in the deleted point cloud data sequence 106. The target point cloud set in the second point cloud data of the aforementioned point cloud data sequence 103 is deleted, resulting in the deleted second point cloud data 1062 in the deleted point cloud data sequence 106. The target point cloud set in the third point cloud data of the aforementioned point cloud data sequence 103 is deleted, resulting in the deleted third point cloud data 1063 in the deleted point cloud data sequence 106.

[0042] It should be noted that the aforementioned electronic device 101 can be either hardware or software. When the electronic device is hardware, it can be implemented as a distributed cluster consisting of multiple servers or terminal devices, or as a single server or a single terminal device. When the electronic device is software, it can be installed in the hardware devices listed above. It can be implemented as, for example, multiple software programs or software modules used to provide distributed services, or as a single software program or software module. No specific limitations are made here.

[0043] It should be understood that Figure 1 and Figure 2 The number of electronic devices shown is merely illustrative. Any number of electronic devices can be used depending on the implementation requirements.

[0044] Continue to refer to Figure 3 The diagram illustrates a flow 300 of some embodiments of a data processing method according to the present disclosure. This data processing method includes the following steps:

[0045] Step 301: Obtain the point cloud data sequence used to construct the point cloud map.

[0046] In some embodiments, the entity executing the above data processing method (e.g. Figure 1 The electronic device shown can acquire point cloud data sequences for constructing point cloud maps via wired or wireless connections. The aforementioned point cloud data sequences include point cloud data sequences. The aforementioned point cloud map can be a high-precision point cloud map. In engineering, the set of point data on the surface of a product obtained through relevant measuring instruments can be called a point cloud. Point clouds obtained using a 3D coordinate measuring machine typically have a smaller number of points and larger spacing between points, and are called sparse point clouds. Point clouds obtained using a 3D laser scanner or photogrammetric scanner have a larger number of points and are denser, and are called dense point clouds. Point clouds obtained based on laser measurement principles include 3D coordinates (XYZ) and laser reflection intensity (Intensity). Point clouds obtained based on photogrammetry principles can include 3D coordinates (XYZ) and color information (RGB). Point clouds obtained by combining laser measurement and photogrammetry principles can include 3D coordinates (XYZ), laser reflection intensity (Intensity), and color information (RGB). Point cloud formats can include, but are not limited to: *.pts; *.asc; *.dat; *.stl; *.imw; *.xyz. The two adjacent frames of point cloud data in the above point cloud data sequence may have a temporal correlation.

[0047] It should be noted that the aforementioned wireless connection methods may include, but are not limited to, 3G / 4G connection, WiFi connection, Bluetooth connection, WiMAX connection, Zigbee connection, UWB (ultra wideband) connection, and other currently known or future wireless connection methods.

[0048] Step 302: Generate a list of object recognition results corresponding to the above point cloud data sequence.

[0049] In some embodiments, the executing entity can generate a list of object recognition results corresponding to the point cloud data sequence. This list includes information on each identified object and a recognition result for each object. The recognition result may indicate whether each object was detected in each frame of point cloud data. The recognition result may also include time information indicating when each object was detected in each frame of point cloud data.

[0050] As an example, the above list of object recognition results can be as follows: Figure 1 and Figure 2 As shown, there is a one-to-one correspondence between the recognition results of each object information and the point cloud data sequence. Specifically, by detecting the presence of each object information in each frame of point cloud data, the recognition result of each object information can be obtained. Each object information can correspond to a point cloud data sequence, thus, each object information can correspond to multiple recognition results.

[0051] It should be noted that the number of recognition results corresponding to each object information can be the same as or less than the number of point cloud data sequences.

[0052] As an example, the aforementioned execution entity can utilize various deep learning models for object recognition to identify object information present in each frame of point cloud data in a point cloud data sequence. Each object information corresponds to one recognition result, and each frame of point cloud data corresponds to multiple recognition results. Thus, each object information corresponds to multiple recognition results. Furthermore, by constructing a list of object information and each recognition result, a list of object recognition results can be obtained.

[0053] Step 303: Delete the target object information and the corresponding recognition results in the above object recognition result list to obtain the processed object recognition result list.

[0054] In some embodiments, the execution entity may delete target object information and corresponding recognition results from the object recognition result list to obtain a processed object recognition result list. For example, the target object information may be the object information identified the least frequently in each frame of point cloud data.

[0055] In some optional implementations of certain embodiments, the recognition result includes: the number of times object information is recognized and the recognition scores for each object information. Specifically, at least one frame of point cloud data in the point cloud data sequence can be identified as containing target object information. For the point cloud data sequence, the number of times the object information is recognized can be the number of point cloud data points in the sequence in which object information can be identified. Whether each object information exists in each frame of point cloud data can be determined using various deep learning models for object recognition. The output of the deep learning model can be a recognition score, which characterizes the probability of each object information existing in each frame of point cloud data.

[0056] The above-mentioned deletion process of target object information and corresponding recognition results in the object recognition result list to obtain the processed object recognition result list may include the following steps:

[0057] The first step is to determine the number of times each object in the object recognition result list is recognized and the highest recognition score among all the corresponding recognition scores, based on the aforementioned object recognition result list. As an example, the executing entity can use a relevant script to compile the number of times each object in the object recognition result list is recognized and the highest recognition score among all the corresponding recognition scores.

[0058] The second step involves removing target object information whose recognition frequency is less than a first threshold and / or whose highest recognition score is less than a second threshold, along with the corresponding recognition results, from the object recognition result list. This yields a processed object recognition result list. The first and second thresholds can be preset.

[0059] Step 304: Based on the object recognition result list processed above, the target point cloud set in each frame of point cloud data in the above point cloud data sequence is deleted to generate deleted point cloud data, thus obtaining the deleted point cloud data sequence.

[0060] In some embodiments, the execution entity may, based on the processed object recognition result list, perform a deletion process on the target point cloud set in each frame of point cloud data in the point cloud data sequence to generate deleted point cloud data, thus obtaining a deleted point cloud data sequence. The target point cloud set represents a moving obstacle object in each frame of point cloud data.

[0061] As an example, each recognition result corresponding to each object information also includes: the position information of the aforementioned object in the corresponding point cloud data. Based on the position information of each object in each frame of point cloud data, the target point cloud set corresponding to the position information in each frame of point cloud data can be selectively deleted to generate deleted point cloud data, resulting in a deleted point cloud data sequence.

[0062] In some optional implementations of certain embodiments, the process of deleting target point clouds from each frame of point cloud data in the point cloud data sequence based on the processed object recognition result list to generate deleted point cloud data, resulting in a deleted point cloud data sequence, may include the following steps:

[0063] The first step is to determine the occurrence time of each object based on the processed object recognition result list. This occurrence time can be the time when each object first appears in each frame of point cloud data during object recognition. For example, the executing entity can determine the occurrence time of each object from the processed object recognition result list using a data query method.

[0064] The second step involves performing the following operations on each frame of point cloud data, based on the forward order of the aforementioned point cloud data sequence:

[0065] The first sub-step is to determine the time difference between the time corresponding to each frame of point cloud data and the time when each object information appears in each object information.

[0066] The second sub-step involves superimposing the bounding boxes corresponding to each object's information onto the corresponding positions in each frame of point cloud data, based on the aforementioned time difference. The bounding box corresponding to each object's information can be the smallest box that encloses the object.

[0067] As an example, the shape of the bounding box can be a cuboid. As an example, in response to the time difference being less than or equal to a preset threshold, the bounding box corresponding to each object information is superimposed onto the corresponding position in each frame of point cloud data.

[0068] The third sub-step is to determine whether there is a point cloud in at least one bounding box in each frame of point cloud data.

[0069] As an example, the aforementioned execution entity can determine whether there is a point cloud in at least one bounding box in each frame of point cloud data by querying.

[0070] As another example, the aforementioned execution entity can determine whether there is a point cloud in at least one bounding box in each frame of point cloud data using the following formula.

[0071]

[0072] Among them, P box It can be a point cloud P lidar The points are projected onto the bounding box coordinate system. This bounding box coordinate system can be a pre-built coordinate system based on the bounding box. The position of the bounding box is defined as `position`. `yaw` represents the orientation information of the point cloud in the bounding box coordinate system. The width of the bounding box can be `width`. The height of the bounding box can be `height`.

[0073] The fourth sub-step, in response to the confirmation of existence, adds the geometric center information of the bounding box associated with the object information from the corresponding point cloud data to each recognition result of each object information in the above object recognition result list, thus obtaining the object recognition result list after addition. It should be noted that there may be multiple bounding boxes in each frame of point cloud data, therefore, the bounding box associated with the object information should be added to each recognition result of each object information in the object recognition result list.

[0074] Optionally, based on the processed object recognition result list, the target point cloud set in each frame of the point cloud data sequence is deleted to generate deleted point cloud data, resulting in a deleted point cloud data sequence. The method further includes:

[0075] The first step is to perform the following operations on each frame of point cloud data, following the reverse order of the point cloud data sequence described above:

[0076] The first sub-step is to determine the time difference between the time corresponding to each frame of point cloud data and the time when each object information appears in each object information.

[0077] The second sub-step involves superimposing the bounding box corresponding to each object information onto the corresponding position in each frame of point cloud data, based on the aforementioned time difference.

[0078] The third sub-step is to determine whether there is a point cloud in at least one bounding box in each frame of point cloud data.

[0079] The fourth sub-step, in response to the confirmation of existence, adds the geometric center information of the bounding box associated with the object from the corresponding point cloud data to each recognition result of the object information in the above-added object recognition result list.

[0080] Optionally, the fourth sub-step further includes: based on the geometric center information of the bounding box corresponding to each frame of point cloud data and the volume information of the bounding box, deleting the target point cloud set in each frame of point cloud data in the point cloud data sequence to generate deleted point cloud data, thus obtaining a deleted point cloud data sequence. As an example, the execution entity can determine a spatial range based on the geometric center information of the bounding box in each frame of point cloud data and the volume information of the bounding box, and then delete the target point cloud set in each frame of point cloud data within the spatial range to generate deleted point cloud data, thus obtaining a deleted point cloud data sequence. This achieves the effect of removing moving obstacles.

[0081] In some optional implementations of certain embodiments, the deleted point cloud data sequence is stored for subsequent construction of the point cloud map. As an example, the executing entity may store the deleted point cloud data sequence in a database used to construct the point cloud map.

[0082] The above embodiments of this disclosure have the following beneficial effects: The data processing methods of some embodiments of this disclosure can quickly and efficiently delete target point clouds associated with moving obstacle objects in each frame of point cloud data in a point cloud data sequence. Specifically, when identifying moving obstacle objects based on a deep learning model, the accuracy and recall rate of identification often cannot reach 100%, so there may be erroneous deletions or remaining movable obstacle objects. Especially for moving vehicles, even if a few frames are left, obvious ghosting will still occur, and the noise filtering effect is not good. Based on this, the data processing methods of some embodiments of this disclosure can first obtain a point cloud data sequence for constructing a point cloud map. Then, an object recognition result list corresponding to the above point cloud data sequence is generated. The above object recognition result list includes the information of each identified object and the recognition result of each object information. Here, by generating the object recognition result list, the objects associated with the point cloud data sequence can be effectively identified. This can be used for the subsequent determination of moving obstacle objects. Furthermore, the target object information and the recognition results corresponding to the target object information in the above object recognition result list are deleted to obtain the processed object recognition result list. Here, deleting the target object information and its corresponding recognition results from the object recognition result list effectively excludes objects that are not moving obstacles. Finally, based on the processed object recognition result list, the target point cloud set in each frame of the point cloud data sequence is deleted to generate a deleted point cloud data sequence. Here, the target point cloud set represents the moving obstacle object in each frame of the point cloud data. Therefore, this data processing method can quickly and efficiently delete the target point cloud set associated with moving obstacle objects from each frame of the point cloud data sequence.

[0083] Further reference Figure 4 The diagram illustrates a flow 400 of another embodiment of the data processing method according to this disclosure. This data processing method includes the following steps:

[0084] Step 401: Obtain the point cloud data sequence used to construct the point cloud map.

[0085] Step 402: In response to the fact that the target point cloud data in the above point cloud data sequence is not the first frame of point cloud data in the above point cloud data sequence, the following object recognition result list generation steps are performed:

[0086] The first sub-step is to determine at least one recognition result of the aforementioned target point cloud data.

[0087] In some embodiments, the executing entity (e.g. Figure 1 or Figure 2The electronic device shown can determine at least one recognition result of the aforementioned target point cloud data. As an example, the executing entity can utilize various deep learning models for object recognition to identify object information present in each frame of point cloud data in a point cloud data sequence, thereby determining at least one recognition result of the aforementioned target point cloud data.

[0088] The second sub-step involves integrating at least one of the above recognition results into the initial object recognition result list based on preset conditions, thereby obtaining the integrated initial object recognition result list.

[0089] In some embodiments, the executing entity can integrate at least one recognition result into an initial object recognition result list according to preset conditions to obtain an integrated initial object recognition result list. The initial object recognition result list is related to the point cloud data sequence preceding the target point cloud data in the point cloud data sequence. As an example, the executing entity can first calculate the distance between each recognition result and all recognition results in the initial object recognition result list to generate a distance value set group. Then, it determines the minimum distance value in the distance value set group corresponding to each recognition result. Furthermore, it determines whether the minimum distance value is less than or equal to a preset threshold. If the minimum distance value is less than or equal to the preset threshold, each recognition result is added to the respective recognition results of the target object information in the initial object recognition result list corresponding to the minimum distance value, resulting in an integrated initial object recognition result list. The minimum distance value of each recognition result corresponds to the target recognition result in the initial object recognition result list. The target recognition result information exists in the respective recognition results within the target object information.

[0090] The third sub-step is to determine the initial object recognition result list after integration as the initial object recognition result list.

[0091] In some embodiments, the aforementioned executing entity may determine the integrated initial object recognition result list as the initial object recognition result list.

[0092] The fourth sub-step is to determine whether the target point cloud data exists in the next frame of point cloud data.

[0093] In some embodiments, the execution entity can determine whether the target point cloud data exists in the next frame of point cloud data. For example, the execution entity can determine whether the target point cloud data exists in the next frame of point cloud data by querying.

[0094] The fifth sub-step is to determine the above-mentioned initial object recognition result list as the above-mentioned object recognition result list in response to the absence of the object.

[0095] In some embodiments, in response to the absence of an object, the execution entity may determine the initial object recognition result list as the object recognition result list.

[0096] Step 403: In response to the existence, determine the next frame of point cloud data of the above target point cloud data as the above target point cloud data, and continue to execute the above object recognition result list generation step.

[0097] In some embodiments, in response to the existence of the target point cloud data, the execution entity may determine the next frame of point cloud data as the target point cloud data and continue to execute the object recognition result list generation step.

[0098] In some optional implementations of certain embodiments, the above steps further include:

[0099] In response to the fact that the target point cloud data is the first frame of point cloud data in the point cloud data sequence, at least one object recognition result corresponding to the first frame of point cloud data is filled into a pre-established empty initial object recognition result list, and the next frame of point cloud data of the first frame of point cloud data is determined as the target point cloud data, and the object recognition result list generation step is continued.

[0100] Step 404: Delete the target object information and the corresponding recognition results in the above object recognition result list to obtain the processed object recognition result list.

[0101] Step 405: Based on the object recognition result list processed above, the target point cloud set in each frame of point cloud data in the above point cloud data sequence is deleted to generate deleted point cloud data, thus obtaining the deleted point cloud data sequence.

[0102] In some embodiments, the specific implementation of steps 401, 403-405 and their resulting technical effects can be found in [reference needed]. Figure 3 Steps 301 and 303-304 in the corresponding embodiments will not be repeated here.

[0103] from Figure 4 It can be seen from this that, with Figure 3 Compared to the description of some corresponding embodiments, Figure 4 The data processing method flowchart 400 in some corresponding embodiments further emphasizes the specific steps for constructing the object recognition result list. Therefore, the schemes described in these embodiments more accurately and efficiently remove target point clouds associated with moving obstacle object information from each frame of point cloud data in the point cloud data sequence by precisely and effectively constructing the object recognition result list.

[0104] Further reference Figure 5As an implementation of the methods shown in the above figures, this disclosure provides some embodiments of a data processing apparatus, which are similar to... Figure 2 Corresponding to the method embodiments shown, the device can be specifically applied to various electronic devices.

[0105] like Figure 5 As shown, a data processing apparatus 500 includes: an acquisition unit 501, a generation unit 502, a first processing unit 503, and a second processing unit 504. The acquisition unit 501 is configured to acquire a point cloud data sequence for constructing a point cloud map. The generation unit 502 is configured to generate an object recognition result list corresponding to the point cloud data sequence, wherein the object recognition result list includes information on each identified object and a recognition result for each object. The first processing unit 503 is configured to delete target object information and corresponding recognition results from the object recognition result list to obtain a processed object recognition result list. The second processing unit 504 is configured to, based on the processed object recognition result list, delete target point cloud sets from each frame of point cloud data in the point cloud data sequence to generate deleted point cloud data, resulting in a deleted point cloud data sequence, wherein the target point cloud sets represent moving obstacle objects in each frame of point cloud data.

[0106] In some optional implementations of certain embodiments, the generation unit 502 in the data processing apparatus 500 may be further configured to: in response to the fact that the target point cloud data in the point cloud data sequence is not the first frame point cloud data in the point cloud data sequence, perform the following object recognition result list generation steps: determine at least one recognition result of the target point cloud data; integrate the at least one recognition result into an initial object recognition result list according to preset conditions to obtain an integrated initial object recognition result list, wherein the initial object recognition result list is related to the point cloud data sequence preceding the target point cloud data in the point cloud data sequence; determine the integrated initial object recognition result list as the initial object recognition result list; determine whether the target point cloud data has a next frame point cloud data; in response to the absence of a next frame point cloud data, determine the initial object recognition result list as the object recognition result list.

[0107] In some optional implementations of some embodiments, the generation unit 502 in the data processing device 500 may be further configured to: in response to the existence of the target point cloud data, determine the next frame of point cloud data as the target point cloud data, and continue to execute the object recognition result list generation step.

[0108] In some optional implementations of certain embodiments, the apparatus 500 further includes a filling unit (not shown in the figure). The filling unit can be configured to: in response to the target point cloud data being the first frame of point cloud data in the point cloud data sequence, fill at least one object recognition result corresponding to the first frame of point cloud data into a pre-established, empty initial object recognition result list, and determine the next frame of point cloud data of the first frame of point cloud data as the target point cloud data, and continue executing the object recognition result list generation step.

[0109] In some optional implementations of some embodiments, the above-mentioned recognition result includes: the number of times an object is recognized and the recognition scores of each object; and the first processing unit 503 in the data processing device 500 may be further configured to: determine the number of times each object is recognized and the highest recognition score among the corresponding recognition scores according to the above-mentioned object recognition result list; remove target objects whose number of recognitions is less than a first threshold and / or whose highest recognition score is less than a second threshold and the recognition results corresponding to the target objects from the above-mentioned object recognition result list to obtain the processed object recognition result list.

[0110] In some optional implementations of certain embodiments, the second processing unit 504 in the data processing device 500 may be further configured to: determine the occurrence time of each object information according to the processed object recognition result list; and perform the following operations on each frame of point cloud data according to the forward order of the point cloud data sequence: determine the time difference between the time corresponding to each frame of point cloud data and the occurrence time of each object information in the object information; based on the time difference, superimpose the bounding box corresponding to each object information onto the corresponding position in each frame of point cloud data; determine whether there is a point cloud in at least one bounding box in each frame of point cloud data; and in response to the determination that there is, add the geometric center information of the bounding box associated with the object information in the corresponding point cloud data to each recognition result of each object information in the object recognition result list, thereby obtaining the added object recognition result list.

[0111] In some optional implementations of certain embodiments, the second processing unit 504 in the data processing apparatus 500 may be further configured to: perform the following operations on each frame of point cloud data according to the reverse order of the point cloud data sequence: determine the time difference between the time corresponding to each frame of point cloud data and the time of occurrence of each object information in each object information; according to the time difference, superimpose the bounding box corresponding to each object information onto the corresponding position in each frame of point cloud data; determine whether there is a point cloud in at least one bounding box in each frame of point cloud data; in response to determining that there is, add the geometric center information of the bounding box associated with the object information in the corresponding point cloud data to each recognition result of each object information in the added object recognition result list.

[0112] In some optional implementations of some embodiments, the second processing unit 504 in the data processing device 500 may be further configured to: perform deletion processing on the target point cloud set in each frame of point cloud data in the point cloud data sequence according to the geometric center information of the bounding box corresponding to each frame of point cloud data and the volume information of the bounding box to generate deleted point cloud data, thereby obtaining the deleted point cloud data sequence.

[0113] In some optional implementations of certain embodiments, the apparatus 500 further includes a storage unit (not shown in the figure). The storage unit can be configured to store the deleted point cloud data sequence for subsequent construction of the point cloud map.

[0114] It is understandable that the units described in the device 500 are related to the reference. Figure 3 The steps in the described method correspond accordingly. Therefore, the operations, features, and beneficial effects described above for the method also apply to the device 500 and the units contained therein, and will not be repeated here.

[0115] The following is for reference. Figure 6 It illustrates electronic devices suitable for implementing some embodiments of this disclosure (e.g., Figure 1 or Figure 2 A schematic diagram of the structure of the electronic device (600) in the device. Figure 6 The electronic device shown is merely an example and should not be construed as limiting the functionality and scope of the embodiments of this disclosure.

[0116] like Figure 6As shown, electronic device 600 may include a processing device (e.g., a central processing unit, a graphics processor, etc.) 601, which can perform various appropriate actions and processes according to a program stored in read-only memory (ROM) 602 or a program loaded from storage device 608 into random access memory (RAM) 603. RAM 603 also stores various programs and data required for the operation of electronic device 600. Processing device 601, ROM 602, and RAM 603 are interconnected via bus 604. Input / output (I / O) interface 605 is also connected to bus 604.

[0117] Typically, the following devices can be connected to I / O interface 605: input devices 606 including, for example, touchscreens, touchpads, keyboards, mice, cameras, microphones, accelerometers, gyroscopes, etc.; output devices 607 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; storage devices 608 including, for example, magnetic tapes, hard disks, etc.; and communication devices 609. Communication device 609 allows electronic device 600 to communicate wirelessly or wiredly with other devices to exchange data. Although Figure 6 An electronic device 600 with various devices is shown; however, it should be understood that it is not required to implement or possess all of the devices shown. More or fewer devices may be implemented or possessed alternatively. Figure 6 Each box shown can represent a device or multiple devices as needed.

[0118] In particular, according to some embodiments of this disclosure, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, some embodiments of this disclosure include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device 609, or installed from a storage device 608, or installed from a ROM 602. When the computer program is executed by the processing device 601, it performs the functions defined above in the methods of some embodiments of this disclosure.

[0119] It should be noted that, in some embodiments of this disclosure, the computer-readable medium described above may be a computer-readable signal medium or a computer-readable storage medium, or any combination thereof. A computer-readable storage medium may be, for example,—but not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In some embodiments of this disclosure, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In some embodiments of this disclosure, a computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium can be any computer-readable medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to: wires, optical fibers, RF (radio frequency), etc., or any suitable combination thereof.

[0120] In some implementations, clients and servers can communicate using any currently known or future-developed network protocol such as HTTP (Hypertext Transfer Protocol) and can interconnect with digital data communication (e.g., communication networks) of any form or medium. Examples of communication networks include local area networks (“LANs”), wide area networks (“WANs”), the Internet (e.g., the Internet of Things), and peer-to-peer networks (e.g., ad hoc peer-to-peer networks), as well as any currently known or future-developed networks.

[0121] The aforementioned computer-readable medium may be included in the aforementioned electronic device; or it may exist independently and not assembled into the electronic device. The aforementioned computer-readable medium carries one or more programs that, when executed by the electronic device, cause the electronic device to: acquire a point cloud data sequence for constructing a point cloud map; generate an object recognition result list corresponding to the aforementioned point cloud data sequence, wherein the object recognition result list includes information on each identified object and a recognition result for each object; delete target object information and corresponding recognition results from the object recognition result list to obtain a processed object recognition result list; and, based on the processed object recognition result list, delete target point cloud sets from each frame of point cloud data in the aforementioned point cloud data sequence to generate deleted point cloud data, obtaining a deleted point cloud data sequence, wherein the target point cloud sets represent moving obstacle objects in each frame of point cloud data.

[0122] Computer program code for performing operations of some embodiments of this disclosure can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0123] 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 disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0124] The units described in some embodiments of this disclosure can be implemented in software or hardware. The described units can also be housed in a processor; for example, a processor may be described as including an acquisition unit, a generation unit, a first processing unit, and a second processing unit. The names of these units do not necessarily limit the specific unit; for example, an acquisition unit may also be described as "a unit for acquiring point cloud data sequences for constructing a point cloud map."

[0125] The functions described above in this document can be performed, at least in part, by one or more hardware logic components. For example, exemplary types of hardware logic components that can be used, without limitation, include: Field Programmable Gate Arrays (FPGAs), Application-Specific Integrated Circuits (ASICs), Application Standard Products (ASSPs), System-on-Chip (SoCs), Complex Programmable Logic Devices (CPLDs), and so on.

[0126] The above description is merely a selection of preferred embodiments of this disclosure and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of the invention involved in the embodiments of this disclosure is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the above-described inventive concept. For example, technical solutions formed by substituting the above-described features with (but not limited to) technical features with similar functions disclosed in the embodiments of this disclosure.

Claims

1. A data processing method, comprising: Obtain the point cloud data sequence used to construct the point cloud map; Generate an object recognition result list corresponding to the point cloud data sequence, wherein the object recognition result list includes information on each recognized object and recognition results for each object information, and the recognition results include information on whether each object information was detected in each frame of point cloud data and the time when each object information was detected in each frame of point cloud data. The target object information and each recognition result corresponding to the target object information in the object recognition result list are deleted. The target object information is the object information that is recognized the least in each frame of point cloud data. Based on the order of the point cloud data sequence, the following operations are performed on each frame of point cloud data: determining the time difference between the time corresponding to each frame of point cloud data and the time of occurrence of each object information in each object information; according to the time difference, superimposing the bounding box corresponding to each object information onto the corresponding position in each frame of point cloud data; determining whether at least one bounding box in each frame of point cloud data contains a point cloud; in response to determining that it exists, adding the geometric center information of the bounding box associated with the object information in the corresponding point cloud data to each recognition result of each object information in the object recognition result list, respectively, to obtain the processed object recognition result list; Based on the processed object recognition result list, the target point cloud set in each frame of point cloud data in the point cloud data sequence is deleted to generate deleted point cloud data, resulting in a deleted point cloud data sequence, wherein the target point cloud set represents the moving obstacle object in each frame of point cloud data.

2. The method according to claim 1, wherein, The generation of the object recognition result list corresponding to the point cloud data sequence includes: In response to the fact that the target point cloud data in the point cloud data sequence is not the first frame of point cloud data in the point cloud data sequence, the following object recognition result list generation steps are performed: Determine at least one recognition result of the target point cloud data; The at least one recognition result is integrated into the initial object recognition result list according to preset conditions to obtain the integrated initial object recognition result list, wherein the initial object recognition result list is related to the point cloud data sequence in the point cloud data sequence that precedes the target point cloud data; The integrated initial object recognition result list is determined as the initial object recognition result list; Determine whether the target point cloud data contains the next frame of point cloud data; In response to the absence of an object, the initial object recognition result list is determined as the object recognition result list.

3. The method according to claim 2, wherein, The method further includes: In response to the existence, the next frame of point cloud data of the target point cloud data is determined as the target point cloud data, and the object recognition result list generation step is continued.

4. The method according to claim 2, wherein, The method further includes: In response to the target point cloud data being the first frame of point cloud data in the point cloud data sequence, at least one object recognition result corresponding to the first frame of point cloud data is written into a pre-established, empty initial object recognition result list, and the next frame of point cloud data of the first frame of point cloud data is determined as the target point cloud data, and the object recognition result list generation step is continued.

5. The method according to claim 2, wherein, The recognition results include: the number of times the object information was recognized and the recognition scores for each aspect of the object information; and The process of deleting target object information and corresponding recognition results from the object recognition result list to obtain a processed object recognition result list includes: Based on the object recognition result list, determine the number of times each object information is recognized and the highest recognition score among the corresponding recognition scores; Remove the target object information whose recognition count is less than the first threshold and / or whose highest recognition score is less than the second threshold, and remove each recognition result corresponding to the target object information from the object recognition result list to obtain the processed object recognition result list.

6. The method according to claim 1, wherein, The step involves deleting target point clouds from each frame of point cloud data in the point cloud data sequence based on the processed object recognition result list to generate deleted point cloud data, resulting in a deleted point cloud data sequence, including: Based on the processed list of object recognition results, determine the occurrence time of each object's information; The operation is performed on each frame of point cloud data according to the forward order of the point cloud data sequence.

7. The method according to claim 6, wherein, The method further includes: Based on the reverse order of the point cloud data sequence, the following operations are performed on each frame of point cloud data: Determine the time difference between the time corresponding to each frame of point cloud data and the time when each object information appears in each object information; Based on the time difference, the bounding box corresponding to each object information is superimposed onto the corresponding position in each frame of point cloud data; Determine whether at least one bounding box in each frame of point cloud data contains a point cloud; In response to the determination of existence, the geometric center information of the bounding box associated with the object information in the corresponding point cloud data is added to each recognition result of each object information in the added object recognition result list.

8. The method according to claim 7, wherein, The method further includes: Based on the geometric center information of the bounding box corresponding to each frame of point cloud data and the volume information of the bounding box, the target point cloud set in each frame of point cloud data in the point cloud data sequence is deleted to generate deleted point cloud data, thus obtaining the deleted point cloud data sequence.

9. The method according to claim 1, wherein, The method further includes: The deleted point cloud data sequence is stored for subsequent construction of the point cloud map.

10. A data processing apparatus, comprising: The acquisition unit is configured to acquire a sequence of point cloud data used to construct a point cloud map; The generation unit is configured to generate an object recognition result list corresponding to the point cloud data sequence, wherein the object recognition result list includes information on each recognized object and a recognition result for each object information, and the recognition result includes information on whether each object information is detected in each frame of point cloud data and the time when each object information is detected in each frame of point cloud data. The first processing unit is configured to perform deletion processing on the target object information and each recognition result corresponding to the target object information in the object recognition result list; and to perform the following operations on each frame of point cloud data according to the order of the point cloud data sequence: determine the time difference between the time corresponding to each frame of point cloud data and the time of occurrence of each object information in each object information; according to the time difference, superimpose the bounding box corresponding to each object information onto the corresponding position in each frame of point cloud data; determine whether there is a point cloud in at least one bounding box in each frame of point cloud data; in response to the determination of existence, add the geometric center information of the bounding box associated with the object information in the corresponding point cloud data to each recognition result of each object information in the object recognition result list, to obtain the processed object recognition result list, wherein the target object information is the object information that is recognized the least in each frame of point cloud data; The second processing unit is configured to perform a deletion process on the target point cloud set in each frame of point cloud data in the point cloud data sequence according to the processed object recognition result list to generate deleted point cloud data, thereby obtaining a deleted point cloud data sequence, wherein the target point cloud set represents the moving obstacle object in each frame of point cloud data.

11. An electronic device, comprising: One or more processors; Storage device, on which one or more programs are stored, When the one or more programs are executed by the one or more processors, the one or more processors implement the method as described in any one of claims 1-9.

12. A computer-readable medium having a computer program stored thereon, wherein, When the program is executed by the processor, it implements the method as described in any one of claims 1-9.

Images