Monocular perception data processing method, device, electronic device and storage medium

Abstract

The present application relates to a monocular perception data processing method, device, electronic device and storage medium, including: obtaining a monocular image captured by a monocular camera; converting the monocular image into point cloud data; converting the point cloud data into polar coordinate data in a two-dimensional polar coordinate system; wherein the two-dimensional polar coordinate system takes the monocular camera as the origin; dividing the two-dimensional polar coordinate system into a plurality of equally divided phase intervals; and determining the target polar coordinate point in each phase interval. The method converts the point cloud data obtained based on the monocular image into polar coordinate data in a two-dimensional polar coordinate system, and by screening the coordinate points divided into each phase interval, retaining the coordinate point with the smallest radius as the target polar coordinate point, downsampling of the point cloud data is achieved, erroneous data in the process of converting the monocular image to the point cloud data is eliminated, and the location information of the obstacle is determined based on the target polar coordinate point retained in each phase interval, thereby improving the recognition accuracy of the obstacle location information.

Description

Technical Field

[0001] The present application relates to the field of data processing technology, and in particular to a monocular perception data processing method, device, electronic device and storage medium. Background Art

[0002] Currently, common low-cost intelligent autonomous mobile robots often use a monocular camera as the main perception sensor to meet various perception requirements in application scenarios, such as basic obstacle detection, boundary detection, and moving object detection. The common process is to obtain images and perform segmentation / detection and other reasoning on the artificial intelligence AI model, and then convert the results into point clouds in 3D space for obstacle avoidance and boundary restrictions. Since monocular cameras lack depth information, the implementation of this process relies on an important prior assumption: obstacles are assumed to be on the horizontal ground.

[0003] Although this assumption is satisfied in most cases in open outdoor scenes, there are still many times when obstacles do not meet this assumption, that is, the obstacles themselves have a certain height, such as standing people, objects with a certain height in the scene such as trees, etc. In this case, the process of converting the above image information to point cloud will produce obvious misestimation, resulting in the inability to accurately determine the location information of the obstacle, thereby interfering with the robot's obstacle avoidance behavior. Summary of the invention

[0004] The present application provides a monocular perception data processing method, device, electronic device and storage medium to solve the technical problem of how to accurately determine obstacle location information based on a monocular camera.

[0005] In a first aspect, the present application provides a method for processing monocular perception data, the method comprising:

[0006] Acquire a monocular image captured by a monocular camera; wherein the monocular image includes at least one obstacle;

[0007] Converting the monocular image into point cloud data;

[0008] Converting the point cloud data into polar coordinate data in a two-dimensional polar coordinate system; wherein the two-dimensional polar coordinate system takes the monocular camera as the origin;

[0009] Dividing the two-dimensional polar coordinate system into a plurality of equally divided phase intervals; wherein the plurality of phase intervals have the origin as a vertex;

[0010] Determine a target polar coordinate point in each phase interval; wherein the target polar coordinate point is a coordinate point with the smallest radius among all interval polar coordinate points in the phase interval;

[0011] The position information of the obstacle is determined according to all the target polar coordinate points.

[0012] Optionally, converting the monocular image into point cloud data comprises:

[0013] estimating depth information of the monocular image;

[0014] Obtaining the intrinsic parameter matrix of the monocular camera;

[0015] The monocular image is converted into the point cloud data according to the depth information and the intrinsic parameter matrix.

[0016] Optionally, estimating depth information of the monocular image includes:

[0017] estimating the depth information of the monocular image using a pre-trained depth estimation model;

[0018] or,

[0019] Acquire a first image captured by the monocular camera; wherein the first image and the monocular image are shot at different angles and / or at different positions; and estimate the depth information of the monocular image based on the first image and the monocular image.

[0020] Optionally, converting the point cloud data into polar coordinate data in a two-dimensional polar coordinate system includes:

[0021] Projecting the point cloud data onto a horizontal plane to obtain horizontal coordinate data on the horizontal plane;

[0022] Convert the horizontal coordinate data into a two-dimensional polar coordinate system to obtain radius data and angle data;

[0023] The radius data and the angle data are used as the polar coordinate data.

[0024] Optionally, the two-dimensional polar coordinate system is divided into a plurality of equally divided phase intervals, including:

[0025] Determine the target number of partition intervals;

[0026] The two-dimensional polar coordinate system is divided into the target number of equally divided phase intervals.

[0027] Optionally, determining a target number of partition intervals includes:

[0028] Get accuracy requirements;

[0029] The target number of divided intervals is determined according to the accuracy requirement and a preset mapping relationship; wherein the mapping relationship is a mapping relationship between the accuracy requirement and the number of divided intervals.

[0030] Optionally, determining a target polar coordinate point in each phase interval includes:

[0031] Determine all interval polar coordinate points contained in the phase interval;

[0032] Obtaining radius information of the interval polar coordinate points;

[0033] Determine the target radius information with the smallest radius among all the radius information;

[0034] The interval polar coordinate point corresponding to the target radius information is used as the target polar coordinate point of the phase interval.

[0035] In a second aspect, the present application provides a monocular perception data processing device, the device comprising:

[0036] An acquisition module, used to acquire a monocular image captured by a monocular camera; wherein the monocular image includes at least one obstacle;

[0037] A first conversion module, used for converting the monocular image into point cloud data;

[0038] A second conversion module, used to convert the point cloud data into polar coordinate data in a two-dimensional polar coordinate system; wherein the two-dimensional polar coordinate system takes the monocular camera as an origin;

[0039] A coordinate system division module, used for dividing the two-dimensional polar coordinate system into a plurality of equally divided phase intervals; wherein the plurality of phase intervals have the origin as a vertex;

[0040] A first determination module is used to determine a target polar coordinate point in each phase interval; wherein the target polar coordinate point is a coordinate point with the smallest radius among all interval polar coordinate points in the phase interval;

[0041] The second determination module is used to determine the position information of the obstacle according to all the target polar coordinate points.

[0042] In a third aspect, the present application provides an electronic device, including a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other via the communication bus;

[0043] Memory, used to store computer programs;

[0044] The processor is used to implement the monocular perception data processing method described in any embodiment of the first aspect when executing the program stored in the memory.

[0045] In a fourth aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the monocular perception data processing method as described in any embodiment of the first aspect is implemented.

[0046] The above technical solution provided by the embodiment of the present application has the following advantages over the prior art: the method provided by the embodiment of the present application obtains a monocular image captured by a monocular camera; wherein the monocular image includes at least one obstacle; the monocular image is converted into point cloud data; the point cloud data is converted into polar coordinate data in a two-dimensional polar coordinate system; wherein the two-dimensional polar coordinate system takes the monocular camera as the origin; the two-dimensional polar coordinate system is divided into a plurality of equally divided phase intervals; wherein the plurality of phase intervals take the origin as the vertex; a target polar coordinate point in each of the phase intervals is determined; wherein the target polar coordinate point is a coordinate point with the smallest radius among all the interval polar coordinate points within the phase interval; and the position information of the obstacle is determined based on all the target polar coordinate points. This method converts the point cloud data obtained based on the monocular image into polar coordinate data in a two-dimensional polar coordinate system. By screening the coordinate points divided into each phase interval, the coordinate point with the smallest radius is retained as the target polar coordinate point, thereby realizing downsampling of the point cloud data, eliminating the erroneous data in the process of converting the monocular image to the point cloud data, and determining the location information of the obstacle based on the target polar coordinate point retained in each phase interval, thereby improving the recognition accuracy of the obstacle location information. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0048] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0049] One or more embodiments are exemplarily described by pictures in the corresponding drawings, and these exemplified descriptions do not constitute limitations on the embodiments. Elements with the same reference numerals in the drawings represent similar elements, and unless otherwise stated, the figures in the drawings do not constitute proportional limitations.

[0050] Figure 1 A system architecture diagram of a monocular perception data processing method provided for one embodiment of the present application;

[0051] Figure 2A flowchart of a monocular perception data processing method provided by one embodiment of the present application;

[0052] Figure 3 A schematic diagram of converting a monocular image into point cloud data;

[0053] Figure 4 A schematic diagram of a two-dimensional polar coordinate system division provided for one embodiment of the present application;

[0054] Figure 5 A schematic diagram of the structure of a monocular perception data processing device provided in one embodiment of the present application;

[0055] Figure 6 A schematic diagram of the structure of an electronic device provided in accordance with an embodiment of the present application. DETAILED DESCRIPTION

[0056] In order to make the purpose, technical solution and advantages of the embodiments of the present application clearer, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0057] The disclosure below provides many different embodiments or examples to realize the different structures of the present application. In order to simplify the disclosure of the present application, the parts and settings of specific examples are described below. Of course, they are only examples, and the purpose is not to limit the present application. In addition, the present application can repeat reference numbers and / or letters in different examples. This repetition is for the purpose of simplification and clarity, and does not itself indicate the relationship between the various embodiments and / or settings discussed.

[0058] In order to solve the technical problem of how to accurately determine the obstacle location information based on a monocular camera in the prior art, the present application provides a monocular perception data processing method, device, electronic device and storage medium, which can eliminate erroneous data in the process of converting monocular images to point cloud data, and determine the obstacle location information based on the target polar coordinate points retained in each phase interval, thereby improving the recognition accuracy of the obstacle location information.

[0059] The first embodiment of the present application provides a monocular perception data processing method, which can be applied to Figure 1The system architecture shown in the figure includes at least a data acquisition module 101 and a data processing module 102, and the data acquisition module 101 and the data processing module 102 establish a communication connection. Specifically, the system architecture can be an intelligent autonomous mobile robot equipped with a monocular camera, such as a lawn mowing robot, a sweeping robot, etc., without limitation.

[0060] Next, based on the system architecture, the monocular perception data processing method is described in detail. Figure 2 , the monocular perception data processing method includes:

[0061] Step 201, obtaining a monocular image captured by a monocular camera; wherein the monocular image includes at least one obstacle.

[0062] The intelligent autonomous mobile robot is equipped with a monocular camera, which is used to identify obstacles in the working scene. When the intelligent autonomous mobile robot is working, the monocular camera is used to collect monocular images.

[0063] Step 202: convert the monocular image into point cloud data.

[0064] In one embodiment, converting a monocular image into point cloud data includes: estimating depth information of the monocular image; obtaining an intrinsic parameter matrix of the monocular camera; and converting the monocular image into point cloud data according to the depth information and the intrinsic parameter matrix.

[0065] In this embodiment, since the monocular image lacks depth information, the intelligent autonomous mobile robot first estimates the depth information of the collected monocular image, and then converts the monocular image into point cloud data with depth information in combination with the intrinsic parameter matrix of the monocular camera.

[0066] In one embodiment, the depth information of a monocular image may be estimated in at least the following two ways:

[0067] The first method is to use a pre-trained depth estimation model to estimate the depth information of a monocular image.

[0068] The monocular training samples may be used to train a depth estimation model, so that after a monocular image is obtained, the depth estimation model may be used to estimate the depth information of the monocular image.

[0069] The second method is to obtain a first image captured by a monocular camera; wherein the first image and the monocular image are shot at different angles and / or different positions; and the depth information of the monocular image is estimated based on the first image and the monocular image.

[0070] The depth information of one of the monocular images can be determined by using multiple monocular images obtained by a monocular camera at different shooting angles or different shooting positions.

[0071] Of course, the point cloud data obtained by the above method is not actually accurate because it is estimated through a monocular image. It can only roughly estimate the distance information of the person or object in the image, and the height information cannot be determined. If the point cloud is projected onto the horizontal plane based on this, when the obstacle has a certain height (such as an upright person, a tree trunk, a table, a chair, etc.), the projected point cloud is only accurate near the camera (corresponding to the part of the obstacle close to the ground), and most of the point cloud is the result of incorrect projection (corresponding to the part of the obstacle that is not close to the ground and has a height). A schematic diagram of converting monocular image point cloud data is shown below: Figure 3 ,in, Figure 3 a in the figure is a monocular image captured by a monocular camera. When the vertical obstacle is converted into an obstacle point cloud, the point cloud is distributed on a horizontal ground. Figure 3 b in which, Figure 3 The perspective of b in Figure 3 The perspective of a in is the same, Figure 3 The c in the figure is a bird's-eye view of the point cloud data, that is, from the perspective of looking down at the horizontal ground, the obstacle point cloud converted from the upright obstacle is actually "lying" on a horizontal ground.

[0072] Step 203, converting the point cloud data into polar coordinate data in a two-dimensional polar coordinate system; wherein the two-dimensional polar coordinate system takes the monocular camera as the origin.

[0073] In one embodiment, point cloud data is converted into polar coordinate data in a two-dimensional polar coordinate system, including: projecting the point cloud data onto a horizontal plane to obtain horizontal coordinate data on the horizontal plane; converting the horizontal coordinate data into a two-dimensional polar coordinate system to obtain radius data and angle data; and using the radius data and angle data as polar coordinate data.

[0074] In this embodiment, for example, in the camera coordinate system, the coordinates of any point in the origin cloud are p (x, y, z), and z is ignored (because z does not affect navigation). It is transformed into a two-dimensional polar coordinate system with the monocular camera as the origin. Point p is transformed into p'(r, theta), where r represents the radius from the point p' to the origin, and theta represents the angle of the point p' in the two-dimensional polar coordinate system. The position of point p' in the two-dimensional polar coordinate system is represented by radius data and angle data.

[0075] Step 204, dividing the two-dimensional polar coordinate system into a plurality of equally divided phase intervals; wherein the plurality of phase intervals have the origin as a vertex.

[0076] In one embodiment, dividing the two-dimensional polar coordinate system into a plurality of equally divided phase intervals includes: determining a target number of divided intervals; and dividing the two-dimensional polar coordinate system into the target number of equally divided phase intervals.

[0077] In this embodiment, the two-dimensional polar coordinate system can be divided into a plurality of equally divided phase intervals, for example, the 180° range in front of the camera is divided into 180 intervals with the same angle with the origin as the vertex. Of course, it can also be divided into 90, 45 or 360 intervals, and the number of interval divisions is not limited. Since there is no image collected in the 180° range behind the camera, it can be not divided, or it can be ignored after being divided, and there is no restriction.

[0078] The target number of partition intervals can be a pre-configured fixed number or determined according to the accuracy requirements.

[0079] In one embodiment, determining the target number of divided intervals includes: obtaining accuracy requirements; determining the target number of divided intervals according to the accuracy requirements and a preset mapping relationship; wherein the mapping relationship is a mapping relationship between the accuracy requirements and the number of divided intervals.

[0080] In this embodiment, the target number of divided intervals can be determined based on the acquired accuracy requirements and the mapping relationship between the preset accuracy requirements and the number of divided intervals, so as to flexibly adjust the number of interval divisions according to the accuracy requirements to ensure that the number of divided intervals meets the accuracy requirements of the user. When the accuracy requirements are high, a larger number of divided intervals can be selected, and when the accuracy requirements are low, a smaller number of divided intervals can be selected. The embodiment of the present application does not impose a specific restriction on the number of divided intervals, but determines it according to the accuracy requirements, taking into account the accuracy while ensuring the speed of data processing, and does not blindly choose high precision or low precision.

[0081] In this embodiment, the schematic diagram of the division of the two-dimensional polar coordinate system is as follows: Figure 4 , the two-dimensional polar coordinate system can be divided into multiple identical phase intervals. For example, the 180° range in front of the camera is divided into multiple intervals with the origin as the vertex. The 180° range behind the camera does not need to be divided because no image is collected. The angle occupied by each phase interval is the same. The polar coordinate points contained in the phase interval can be downsampled to eliminate the erroneous data in the process of converting the monocular image to point cloud data. Therefore, the location information of the obstacle can be determined based on the target polar coordinate points retained in each phase interval, thereby improving the recognition accuracy of the obstacle location information.

[0082] It should be noted that Figure 4 The number of partition intervals in the figure is only an example to illustrate how to partition, and does not represent a limit on the number of partition intervals.

[0083] Step 205, determining a target polar coordinate point in each phase interval; wherein the target polar coordinate point is a coordinate point with the smallest radius among all interval polar coordinate points in the phase interval.

[0084] In one embodiment, determining the target polar coordinate point in each phase interval includes: determining all interval polar coordinate points contained in the phase interval; obtaining radius information of the interval polar coordinate points; determining the target radius information with the smallest radius among all the radius information; and using the interval polar coordinate point corresponding to the target radius information as the target polar coordinate point of the phase interval.

[0085] In this embodiment, see Figure 4 Different phase intervals contain different numbers of interval polar coordinate points. For any phase interval, the coordinate point with the smallest radius among all interval polar coordinate points in the phase interval can be taken as the target polar coordinate point, that is, each phase interval containing interval polar coordinate points only retains one coordinate point, and the polar coordinate points contained in each phase interval are downsampled. It should be understood that the phase interval that does not contain interval polar coordinate points can be processed. Since in the process of converting monocular images to point clouds, the point cloud data obtained by projection is only an accurate point cloud close to the camera (corresponding to the part of the obstacle close to the ground), and most of the point clouds are the result of erroneous projection (corresponding to the part of the obstacle that is not close to the ground and has a height), therefore, retaining the polar coordinate point with the smallest radius can retain the accurate point cloud data and eliminate the inaccurate point cloud data. That is, the point cloud processed by the method of the embodiment of the present application not only retains the core obstacle information (the point cloud mapped from the bottom of the obstacle, with accurate obstacle location information), but also automatically filters out the point cloud obtained by erroneous projection due to the height of the obstacle. At the same time, the number of point clouds is also reduced. For an intelligent autonomous mobile robot that applies this method, it helps to reduce the communication bandwidth pressure between the perception and navigation modules in the robot system and reduce the amount of calculation of the obstacle cost map by the navigation module.

[0086] This embodiment aims at the common process of monocular perception that currently relies on an important prior assumption that "obstacles are assumed to be located on the horizontal ground". When dealing with the problem of perception of obstacles at a certain height, a downsampling algorithm that takes occlusion into consideration is performed on the point cloud converted to the polar coordinate system. This achieves the goal of retaining the core obstacle information and filtering out the point cloud obtained by erroneous projection due to the height of the obstacle. At the same time, the number of point clouds is reduced without affecting obstacle avoidance and boundary restrictions, thereby reducing system data transmission and computing load.

[0087] Step 206: Determine the location information of the obstacle according to all target polar coordinate points.

[0088] This method converts the point cloud data obtained based on the monocular image into polar coordinate data in a two-dimensional polar coordinate system. By screening the coordinate points divided into each phase interval, the coordinate point with the smallest radius is retained as the target polar coordinate point, thereby realizing downsampling of the point cloud data, eliminating the erroneous data in the process of converting the monocular image to the point cloud data, and determining the location information of the obstacle based on the target polar coordinate point retained in each phase interval, thereby improving the recognition accuracy of the obstacle location information.

[0089] Based on the same technical concept, the second embodiment of the present application provides a monocular perception data processing device, such as Figure 5 , the device comprises:

[0090] The acquisition module 501 is used to acquire a monocular image captured by a monocular camera; wherein the monocular image includes at least one obstacle;

[0091] A first conversion module 502, used to convert the monocular image into point cloud data;

[0092] A second conversion module 503 is used to convert the point cloud data into polar coordinate data in a two-dimensional polar coordinate system; wherein the two-dimensional polar coordinate system takes the monocular camera as the origin;

[0093] A coordinate system division module 504 is used to divide the two-dimensional polar coordinate system into a plurality of equally divided phase intervals; wherein the plurality of phase intervals have the origin as a vertex;

[0094] A first determination module 505 is used to determine a target polar coordinate point in each phase interval; wherein the target polar coordinate point is a coordinate point with the smallest radius among all interval polar coordinate points in the phase interval;

[0095] The second determining module 506 is used to determine the position information of the obstacle according to all the target polar coordinate points.

[0096] The device converts point cloud data obtained based on a monocular image into polar coordinate data in a two-dimensional polar coordinate system, screens the coordinate points divided into each phase interval, retains the coordinate point with the smallest radius as the target polar coordinate point, realizes downsampling of the point cloud data, eliminates erroneous data in the process of converting the monocular image to point cloud data, determines the position information of the obstacle based on the target polar coordinate point retained in each phase interval, and improves the recognition accuracy of the obstacle position information.

[0097] like Figure 6 As shown, an embodiment of the present application provides an electronic device, including a processor 111, a communication interface 112, a memory 113 and a communication bus 114, wherein the processor 111, the communication interface 112, and the memory 113 communicate with each other through the communication bus 114.

[0098] Memory 113, used for storing computer programs;

[0099] In one embodiment of the present application, the processor 111 is used to execute the program stored in the memory 113 to implement the monocular perception data processing method provided by any of the above method embodiments, including:

[0100] Acquire a monocular image captured by a monocular camera; wherein the monocular image includes at least one obstacle;

[0101] Converting the monocular image into point cloud data;

[0102] Converting the point cloud data into polar coordinate data in a two-dimensional polar coordinate system; wherein the two-dimensional polar coordinate system takes the monocular camera as the origin;

[0103] Dividing the two-dimensional polar coordinate system into a plurality of equally divided phase intervals; wherein the plurality of phase intervals have the origin as a vertex;

[0104] Determine a target polar coordinate point in each phase interval; wherein the target polar coordinate point is a coordinate point with the smallest radius among all interval polar coordinate points in the phase interval;

[0105] The position information of the obstacle is determined according to all the target polar coordinate points.

[0106] The communication bus mentioned in the above terminal can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. The communication bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, only one thick line is used in the figure, but it does not mean that there is only one bus or one type of bus.

[0107] The communication interface is used for communication between the above terminal and other devices.

[0108] The memory may include a random access memory (RAM) or a non-volatile memory, such as at least one disk memory. Optionally, the memory may also be at least one storage device located away from the aforementioned processor.

[0109] The above-mentioned processor can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

[0110] An embodiment of the present application also provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the monocular perception data processing method provided in any of the aforementioned method embodiments is implemented.

[0111] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0112] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a general hardware platform, and of course, by hardware. Based on this understanding, the above technical solution, in essence, or the part that contributes to the relevant technology, can be embodied in the form of a software product, which can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiment.

[0113] It should be understood that the terms used herein are only for the purpose of describing specific example embodiments and are not intended to be limiting. Unless the context clearly indicates otherwise, the singular forms "one", "an" and "said" as used herein may also be meant to include plural forms. The terms "include", "comprise", "contain", and "have" are inclusive, and therefore specify the existence of stated features, steps, operations, elements and / or parts, but do not exclude the existence or addition of one or more other features, steps, operations, elements, parts, and / or combinations thereof. The method steps, processes, and operations described herein are not interpreted as necessarily requiring them to be performed in the specific order described or illustrated, unless the execution order is clearly indicated. It should also be understood that additional or alternative steps may be used.

[0114] It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application. In the description, the suffixes such as "module", "component" or "unit" used to represent the elements are only used to facilitate the description of the present application and have no specific meaning in themselves. Therefore, "module", "component" or "unit" can be used in a mixed manner.

[0115] The above description is only a specific implementation of the present application, so that those skilled in the art can understand or implement the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest range consistent with the principles and novel features applied for herein.

Claims

1. A monocular perception data processing method, characterized in that: The method comprises: Acquire a monocular image captured by a monocular camera; wherein the monocular image includes at least one obstacle; Converting the monocular image into point cloud data; Converting the point cloud data into polar coordinate data in a two-dimensional polar coordinate system; wherein the two-dimensional polar coordinate system takes the monocular camera as the origin; Dividing the two-dimensional polar coordinate system into a plurality of equally divided phase intervals; wherein the plurality of phase intervals have the origin as a vertex; Determine a target polar coordinate point in each phase interval; wherein the target polar coordinate point is a coordinate point with the smallest radius among all interval polar coordinate points in the phase interval; Determine the position information of the obstacle according to all the target polar coordinate points; Wherein, determining the target polar coordinate point in each phase interval includes: Determine all interval polar coordinate points contained in the phase interval; Obtaining radius information of the interval polar coordinate points; Determine the target radius information with the smallest radius among all the radius information; The interval polar coordinate point corresponding to the target radius information is used as the target polar coordinate point of the phase interval.

2. The method according to claim 1, characterized in that Converting the monocular image into point cloud data includes: estimating depth information of the monocular image; Obtaining the intrinsic parameter matrix of the monocular camera; The monocular image is converted into the point cloud data according to the depth information and the intrinsic parameter matrix.

3. The method according to claim 2, characterized in that Estimating the depth information of the monocular image, comprising: estimating the depth information of the monocular image using a pre-trained depth estimation model; or, Acquire a first image captured by the monocular camera; wherein the first image and the monocular image are shot at different angles and / or at different positions; and estimate the depth information of the monocular image based on the first image and the monocular image.

4. The method according to claim 1, characterized in that: Converting the point cloud data into polar coordinate data in a two-dimensional polar coordinate system includes: Projecting the point cloud data onto a horizontal plane to obtain horizontal coordinate data on the horizontal plane; Convert the horizontal coordinate data into a two-dimensional polar coordinate system to obtain radius data and angle data; The radius data and the angle data are used as the polar coordinate data.

5. The method according to claim 1, characterized in that Dividing the two-dimensional polar coordinate system into a plurality of equally divided phase intervals, including: Determine the target number of partition intervals; The two-dimensional polar coordinate system is divided into phase intervals that are equally divided by the number of targets.

6. The method according to claim 5, characterized in that Determine the target number of partition intervals, including: Get accuracy requirements; The target number of divided intervals is determined according to the accuracy requirement and a preset mapping relationship; wherein the mapping relationship is a mapping relationship between the accuracy requirement and the number of divided intervals.

7. A monocular perception data processing device, characterized in that: The device comprises: An acquisition module, used to acquire a monocular image captured by a monocular camera; wherein the monocular image includes at least one obstacle; A first conversion module, used for converting the monocular image into point cloud data; A second conversion module, used to convert the point cloud data into polar coordinate data in a two-dimensional polar coordinate system; wherein the two-dimensional polar coordinate system takes the monocular camera as an origin; A coordinate system division module, used for dividing the two-dimensional polar coordinate system into a plurality of equally divided phase intervals; wherein the plurality of phase intervals have the origin as a vertex; The first determination module is used to determine the target polar coordinate point in each phase interval; wherein the target polar coordinate point is the coordinate point with the smallest radius among all interval polar coordinate points in the phase interval; determining the target polar coordinate point in each phase interval includes: determining all interval polar coordinate points contained in the phase interval; obtaining radius information of the interval polar coordinate points; determining the target radius information with the smallest radius among all the radius information; and using the interval polar coordinate point corresponding to the target radius information as the target polar coordinate point of the phase interval; The second determination module is used to determine the position information of the obstacle according to all the target polar coordinate points.

8. An electronic device, characterized in that: It includes a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory communicate with each other through the communication bus; Memory, used to store computer programs; The processor is used to implement the monocular perception data processing method described in any one of claims 1 to 6 when executing the program stored in the memory.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the monocular perception data processing method as described in any one of claims 1 to 6 is implemented.

Images