CN117456087A

CN117456087A - Point cloud rasterization method

Info

Publication number: CN117456087A
Application number: CN202311250152.2A
Authority: CN
Inventors: 徐建云; 苗振伟; 刘旻哲; 李坤; 卿泉
Original assignee: Wuzhou Online E Commerce Beijing Co ltd
Current assignee: Zhejiang Cainiao Chuancheng Network Technology Co ltd
Priority date: 2023-09-25
Filing date: 2023-09-25
Publication date: 2024-01-26

Abstract

The embodiment of the specification provides a point cloud rasterizing method, wherein the point cloud rasterizing method comprises the following steps: obtaining a grid diagram to be divided containing a target object; performing grid division processing on the grid graph to be divided according to a preset non-uniform division strategy to obtain a target grid graph, wherein the size of a grid in the target grid graph is in direct proportion to the interval of the grid from the target object; collecting point cloud data aiming at the target object, and constructing a coordinate mapping relation between the point cloud data and grids in the target grid graph; and mapping the point cloud data to the target grid graph based on the coordinate mapping relation.

Description

Point cloud rasterization method

Technical Field

The embodiment of the specification relates to the technical field of automatic driving, in particular to a point cloud rasterization method.

Background

With the vigorous development of the automotive industry, the demand for unmanned driving by the user population is growing. And unmanned technology is increasingly in demand for use in traffic, transportation and unmanned distribution scenarios. In the prior art, in order to realize unmanned driving and improve driving safety, collected point cloud data is usually mapped into a raster pattern by adopting point cloud rasterization processing, so that automatic driving is performed by using the raster pattern mapped with the point cloud data. The three-dimensional point cloud rasterization is mostly implemented by adopting a uniform rasterization mode under a Cartesian system. Although the uniform rasterization method can support unmanned driving, considering the complexity of driving scenes, if the grids are uniformly divided in a wide space, there is a problem in that both resolution and computational resources cannot be fully achieved. Namely, if the resolution is reduced to maintain real-time performance, the aim of saving the computing resources can be achieved, but the accuracy is lower; or to increase resolution to maintain performance, but to greatly increase the consumption of computing resources. There is therefore a need for an effective solution to the above problems.

Disclosure of Invention

In view of this, the present description embodiments provide a point cloud rasterization method. One or more embodiments of the present specification relate to a point cloud rasterizing apparatus, a computing device, a computer-readable storage medium, and a computer program that solve the technical drawbacks of the prior art.

According to a first aspect of embodiments of the present disclosure, there is provided a method for rasterizing a point cloud, including:

obtaining a grid diagram to be divided containing a target object;

performing grid division processing on the grid graph to be divided according to a preset non-uniform division strategy to obtain a target grid graph, wherein the size of a grid in the target grid graph is in direct proportion to the interval of the grid from the target object;

collecting point cloud data aiming at the target object, and constructing a coordinate mapping relation between the point cloud data and grids in the target grid graph;

and mapping the point cloud data to the target grid graph based on the coordinate mapping relation.

According to a second aspect of embodiments of the present disclosure, there is provided another method for rasterizing a point cloud, applied to a vehicle control terminal, including:

acquiring a grid diagram to be divided containing a target vehicle;

Performing grid division processing on the grid graph to be divided according to a preset non-uniform division strategy to obtain a target grid graph, wherein the size of a grid in the target grid graph is in direct proportion to the interval of the grid from the target vehicle;

collecting point cloud data aiming at the target vehicle, and constructing a coordinate mapping relation between the point cloud data and grids in the target grid graph;

and mapping the point cloud data to the target raster pattern based on the coordinate mapping relation, wherein the target raster pattern mapped with the point cloud data is used for automatic driving of the target vehicle.

According to a third aspect of embodiments of the present specification, there is provided a point cloud rasterizing apparatus including:

the acquisition module is configured to acquire a grid graph to be divided containing target objects;

the dividing module is configured to carry out grid division processing on the grid graph to be divided according to a preset non-uniform division strategy to obtain a target grid graph, wherein the size of a grid in the target grid graph is in direct proportion to the interval of the grid from the target object;

the construction module is configured to acquire point cloud data aiming at the target object and construct a coordinate mapping relation between the point cloud data and grids in the target grid graph;

And the mapping module is configured to map the point cloud data to the target grid graph based on the coordinate mapping relation.

According to a fourth aspect of embodiments of the present disclosure, there is provided another point cloud rasterizing apparatus applied to a vehicle control terminal, including:

the grid image acquisition module is configured to acquire a grid image to be divided comprising a target vehicle;

the grid pattern dividing module is configured to perform grid division processing on the grid patterns to be divided according to a preset non-uniform division strategy to obtain target grid patterns, wherein the size of grids in the target grid patterns is in direct proportion to the interval between the grids and the target vehicle;

the construction mapping relation module is configured to acquire point cloud data aiming at the target vehicle and construct a coordinate mapping relation between the point cloud data and grids in the target grid graph;

and the point cloud data mapping module is configured to map the point cloud data to the target grid graph based on the coordinate mapping relation, wherein the target grid graph mapped with the point cloud data is used for automatic driving of the target vehicle.

According to a fifth aspect of embodiments of the present specification, there is provided a computing device comprising:

A memory and a processor;

the memory is configured to store computer-executable instructions that, when executed by the processor, perform the steps of the point cloud rasterization method described above.

According to a sixth aspect of embodiments of the present specification, there is provided a computer readable storage medium storing computer executable instructions which, when executed by a processor, implement the steps of the above-described point cloud rasterization method.

According to a seventh aspect of the embodiments of the present specification, there is provided a computer program, wherein the computer program, when executed in a computer, causes the computer to perform the steps of the above-described point cloud rasterization method.

In the point cloud rasterization method provided by the specification, in order to reduce the consumption of calculation performance while improving the resolution of a region of interest, a near-dense and far-sparse non-uniform rasterization scheme is adopted for raster division. Namely: the method comprises the steps of firstly obtaining a grid diagram to be divided containing a target object, and then carrying out grid division processing on the grid diagram to be divided according to a preset non-uniform division strategy in order to balance the resolution and the consumption of computing resources, so that the resolution corresponding to a region of interest with a relatively close distance to the target object is improved, and meanwhile, less computing resources are used. In the grid dividing process, the grid size in the grid chart is directly proportional to the interval between the grids and the target object, that is, the grid size closer to the target object is smaller, the grid size farther from the target object is larger, the grid dividing process is completed, the near-density and far-thinning process is achieved, the divided non-uniform grids achieve higher processing resolution closer to the target object, more resource consumption, lower processing resolution and less resource consumption in the subsequent process, and therefore the resource consumption is effectively controlled. On the basis, after the point cloud data are acquired, the coordinate position of each grid is not fixed in consideration of the fact that the grids are non-uniform grids, so that the coordinate mapping relation between the point cloud data and the grids in the target grid graph can be constructed firstly, and the mapping relation between the point cloud coordinates in the point cloud data and the grids in the target grid graph can be determined; based on the mapping relation, the point cloud data can be mapped into the target raster pattern, the mapping result is consistent with the position relation in the real scene, and the processing of the downstream task can be realized by realizing the target raster pattern for mapping the point cloud data, so that the aims of balancing the resolution and the resource consumption in the point cloud rasterization processing are fulfilled.

Drawings

FIG. 1 is a schematic diagram of a method for rasterizing a point cloud according to one embodiment of the present disclosure;

FIG. 2 is a flow chart of a method of point cloud rasterization provided in one embodiment of the present disclosure;

FIG. 3 is a schematic diagram of uniform grid division in a method for rasterizing a point cloud according to one embodiment of the present disclosure;

FIG. 4 is a schematic diagram of non-uniform grid division in a method for rasterizing a point cloud according to one embodiment of the present disclosure;

FIG. 5 is a process flow diagram of a method for point cloud rasterization provided in one embodiment of the present disclosure;

FIG. 6 is a flow chart of another method of point cloud rasterization provided by one embodiment of the present disclosure;

fig. 7 is a schematic structural diagram of a point cloud rasterizing apparatus according to an embodiment of the present disclosure;

FIG. 8 is a schematic diagram of another point cloud rasterizing apparatus according to one embodiment of the present disclosure;

FIG. 9 is a block diagram of a computing device provided in one embodiment of the present description.

Detailed Description

In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present description. This description may be embodied in many other forms than described herein and similarly generalized by those skilled in the art to whom this disclosure pertains without departing from the spirit of the disclosure and, therefore, this disclosure is not limited by the specific implementations disclosed below.

The terminology used in the one or more embodiments of the specification is for the purpose of describing particular embodiments only and is not intended to be limiting of the one or more embodiments of the specification. As used in this specification, one or more embodiments and the appended claims, the singular forms "a," "an," and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It should also be understood that the term "and/or" as used in one or more embodiments of the present specification refers to and encompasses any or all possible combinations of one or more of the associated listed items.

It should be understood that, although the terms first, second, etc. may be used in one or more embodiments of this specification to describe various information, these information should not be limited by these terms. These terms are only used to distinguish one type of information from another. For example, a first may also be referred to as a second, and similarly, a second may also be referred to as a first, without departing from the scope of one or more embodiments of the present description. The word "if" as used herein may be interpreted as "at … …" or "at … …" or "responsive to a determination", depending on the context.

Furthermore, it should be noted that, user information (including, but not limited to, user equipment information, user personal information, etc.) and data (including, but not limited to, data for analysis, stored data, presented data, etc.) according to one or more embodiments of the present disclosure are information and data authorized by a user or sufficiently authorized by each party, and the collection, use, and processing of relevant data is required to comply with relevant laws and regulations and standards of relevant countries and regions, and is provided with corresponding operation entries for the user to select authorization or denial.

First, terms related to one or more embodiments of the present specification will be explained.

And (3) point cloud: and a three-dimensional point data set of the actual road surface is acquired through a three-dimensional laser scanning instrument.

A grid: in image editing, a grid is a pixel, meaning that an instruction in an image is converted into a pixel.

In the present specification, a point cloud rasterizing method is provided, and the present specification relates to a point cloud rasterizing apparatus, a computing device, and a computer-readable storage medium, which are described in detail in the following embodiments one by one.

Referring to the schematic diagram shown in fig. 1, in the point cloud rasterization method provided in the present specification, in order to reduce the consumption of computing performance while improving the resolution of the region of interest, a near-dense and far-sparse non-uniform rasterization scheme is used for performing rasterization. Namely: the method comprises the steps of firstly obtaining a grid diagram to be divided containing a target object, and then carrying out grid division processing on the grid diagram to be divided according to a preset non-uniform division strategy in order to balance the resolution and the consumption of computing resources, so that the resolution corresponding to a region of interest with a relatively close distance to the target object is improved, and meanwhile, less computing resources are used. In the grid dividing process, the grid size in the grid chart is directly proportional to the interval between the grids and the target object, that is, the grid size closer to the target object is smaller, the grid size farther from the target object is larger, the grid dividing process is completed, the near-density and far-thinning process is achieved, the divided non-uniform grids achieve higher processing resolution closer to the target object, more resource consumption, lower processing resolution and less resource consumption in the subsequent process, and therefore the resource consumption is effectively controlled. On the basis, after the point cloud data are acquired, the coordinate position of each grid is not fixed in consideration of the fact that the grids are non-uniform grids, so that the coordinate mapping relation between the point cloud data and the grids in the target grid graph can be constructed firstly, and the mapping relation between the point cloud coordinates in the point cloud data and the grids in the target grid graph can be determined; based on the mapping relation, the point cloud data can be mapped into the target raster pattern, the mapping result is consistent with the position relation in the real scene, and the processing of the downstream task can be realized by realizing the target raster pattern for mapping the point cloud data, so that the aims of balancing the resolution and the resource consumption in the point cloud rasterization processing are fulfilled.

Referring to fig. 2, fig. 2 shows a flowchart of a method for rasterizing a point cloud according to an embodiment of the present disclosure, which specifically includes the following steps.

Step S202, a grid diagram to be divided containing target objects is obtained.

The point cloud rasterization method provided by the embodiment can be applied to any unmanned scene, such as unmanned scenes of private cars, unmanned public transport vehicles, unmanned delivery cars and the like; in this embodiment, the unmanned distribution vehicle is taken as an example to describe the point cloud rasterization method, and descriptions in other scenes can be referred to the same or corresponding descriptions in this embodiment, which are not repeated here.

Specifically, the target object specifically refers to an object moving in a current unmanned scene, such as an unmanned delivery vehicle, a sedan, a transport vehicle and the like, and the point cloud rasterization processing needs to be performed on a raster pattern associated with the target object in the current scene, so that safe driving of the target object can be realized. Correspondingly, the grid graph to be divided specifically refers to a two-dimensional graph which is generated for the target object at the current moment and is not subjected to grid division, and the target grid graph is formed after grid division and is used for performing point cloud rasterization processing, so that the target object is driven to run by using the grid graph of the mapping point cloud data in an unmanned driving scene.

Based on this, in order to reduce the consumption of calculation performance while improving the resolution of the region of interest, the raster division may be performed using a near-dense far-sparse non-uniform rasterization scheme. Namely: firstly, a grid diagram to be divided containing a target object is acquired, and then, in order to balance the resolution and the consumption of computing resources, the purpose of improving the resolution corresponding to a region of interest with a relatively close distance to the target object is achieved, meanwhile, fewer computing resources are used, and the grid diagram to be divided can be subjected to grid division processing according to a preset non-uniform division strategy, so that the target grid diagram is obtained. In the grid dividing process, the grid size in the grid chart is directly proportional to the interval between the grids and the target object, that is, the grid size closer to the target object is smaller, the grid size farther from the target object is larger, the grid dividing process is completed, the near-density and far-thinning process is achieved, the divided non-uniform grids achieve higher processing resolution closer to the target object, more resource consumption, lower processing resolution and less resource consumption in the subsequent process, and therefore the resource consumption is effectively controlled.

Further, when the grid graph to be divided containing the target object is obtained, the grid graph to be divided is used for generating a subsequent grid graph and mapping point cloud data, so that unmanned use of the target object is supported, the grid graph to be divided containing the target object can truly reflect the position of the target object, and driving safety can be ensured when unmanned use is ensured, and therefore construction of the grid graph to be divided can be completed by combining attribute information and position information of the target object; in this embodiment, the specific implementation manner is as follows:

loading a target virtual map according to the position information of the target object, and constructing an initial grid map according to the target virtual map; updating the initial raster image based on the position information and the attribute information of the target object, and generating a raster image to be divided containing the target object according to an updating result.

Specifically, the location information specifically refers to a location coordinate corresponding to the current time of the target object in the target virtual map, and is used for locating the location of the target object in the target virtual map. Correspondingly, the target virtual map specifically refers to a map used by the target object in the driving process, and the map has a mapping relation with the driving area of the target object. Correspondingly, the initial raster image specifically refers to a two-dimensional image which is constructed by combining the target virtual map and is not subjected to raster division, and meanwhile, the initial raster image also does not contain target objects. Accordingly, the attribute information specifically refers to description information of basic attributes of the target object, including, but not limited to, size, occupation area, identification, and the like of the target object.

Based on the above, when the grid graph to be constructed including the target object is constructed, the target virtual map corresponding to the current position of the target object can be loaded according to the position information of the target object. On the basis, an initial grid graph can be constructed according to the target virtual map; in order to drive the running of the target object based on the grid graph, the initial grid graph can be updated based on the position information and the attribute information of the target object, so that a mapping object for generating the target object is constructed in the initial grid graph, the target object is displayed in the grid graph in a same way, and the grid graph to be divided containing the target object is generated according to the updating result.

For example, the unmanned delivery vehicle a for express delivery travels on the rightmost lane on the first road, during the traveling process, a virtual map used during the traveling process of the unmanned delivery vehicle a may be loaded according to the current positioning information of the unmanned delivery vehicle a, so as to construct an initial grid map according to the virtual map, and then, the initial grid map is updated by combining the current positioning information of the unmanned delivery vehicle a, the information such as the size and the identification of the unmanned delivery vehicle a, so as to realize that the unmanned delivery vehicle a is embodied in the initial grid map and corresponds to the traveling scene at the current moment, thereby obtaining the grid map to be divided including the unmanned delivery vehicle a for subsequent grid division and point cloud mapping, so as to drive the unmanned delivery vehicle a to travel on the first road safely and in compliance with the traffic rules.

In summary, by adding the target object to the grid graph constructed based on the target virtual map in combination with the position information and the attribute information, the grid graph to be divided which can represent the running state of the target object at the current moment can be obtained, and on the basis, the follow-up processing is performed, so that the authenticity and the running safety of the target object can be ensured.

Step S204, carrying out grid division processing on the grid graph to be divided according to a preset non-uniform division strategy to obtain a target grid graph, wherein the size of a grid in the target grid graph is in direct proportion to the interval of the grid distance from the target object.

Specifically, after the grid graph to be divided including the target object is obtained, further, in order to balance the resolution and the computing resource consumption in two dimensions, the purpose of improving the resolution corresponding to the region of interest with a relatively short distance to the target object while using fewer computing resources is achieved, and a near-dense and far-sparse non-uniform rasterization scheme may be adopted to perform grid division. And performing raster division processing on the raster image to be divided according to a preset non-uniform division strategy, so as to obtain a target raster image. In the grid dividing process, the grid size in the grid chart is directly proportional to the interval between the grids and the target object, that is, the grid size closer to the target object is smaller, the grid size farther from the target object is larger, the grid dividing process is completed, the near-density and far-thinning process is achieved, the divided non-uniform grids achieve higher processing resolution closer to the target object, more resource consumption, lower processing resolution and less resource consumption in the subsequent process, and therefore the resource consumption is effectively controlled.

In practical application, when the near-dense and far-sparse grid distribution is constructed, the grid distribution can be realized by adopting a preset algorithm, and the maximum size and the minimum size of the grid are input into the algorithm, so that the linear distribution is automatically finished by an algorithm model, and the linear distribution is output. The algorithm model may be implemented using a neural network that is fully trained using samples and labels.

The non-uniform division strategy specifically refers to a strategy of dividing grids of a grid graph to be divided according to a near-dense and far-sparse division mode, wherein the near-dense and far-sparse division mode is used for linearly arranging grids from near to far according to small to big, that is, the closer the distance to a target object is, the smaller the distance to the target object is, the farther the distance to the target object is, the larger the grid is, the grid size is arranged according to the linear arrangement, for example, the grids from the target object are arranged according to the distribution from near to far, and adjacent grids are respectively grid 1, grid 2 and grid 3; the size of grid 1 is 0*0, according to the linear arrangement mechanism, the size of grid 2 will be 0.5 x 0.5, and the size of grid 3 will be 1.1 x 1.1; thereby achieving the effect of near-dense and far-sparse. Correspondingly, the target raster image specifically refers to a raster image obtained after raster division processing is performed on the raster image to be divided according to a preset non-uniform division strategy.

In addition, considering that the grid size cannot be infinitely increased, the upper limit of the grid size can be set, and when the grid size obtained in a certain calculation period is larger than or equal to the upper limit of the grid size, all the grid sizes required to be calculated subsequently are set according to the upper limit of the grid size without increasing, so that the actual use requirement is met.

Furthermore, when the raster pattern to be divided is subjected to raster division processing according to a non-uniform division strategy, the raster division processing can be completed by combining the reference raster size and raster distribution parameters in the strategy because the raster is required to be divided according to the proportional relation between the raster size and the distance between the raster and the target object to form a near-dense and far-sparse raster division effect; in this embodiment, the specific implementation manner is as follows:

determining a reference grid size and grid distribution parameters according to a preset non-uniform division strategy; calculating the target grid size of the grids to be divided in the grid queue to be divided according to the reference grid size and the grid distribution parameters; according to the target grid size of grids to be divided in the grid queue to be divided, performing grid division processing on the grid graph to be divided to obtain the target grid graph; the grids to be divided in the grid queue to be divided are ordered according to the interval with the target object, and the calculation priority of the target grid size of the grids to be divided is determined according to the ordering result of the grids to be divided in the grid queue to be divided.

Specifically, the reference grid size specifically refers to the minimum grid size, and correspondingly, the grid distribution parameter specifically refers to the slope of the linear growth of the grid size, and is used for calculating the grid size corresponding to each grid; correspondingly, the grids to be divided specifically refer to grids needing to be subjected to grid size calculation, and the target grid size is the corresponding grid size of the grids to be divided obtained after calculation is completed; correspondingly, the grid queue to be divided specifically means that grids are arranged according to the distance from the target object to obtain a queue, and the grids to be divided in the queue are arranged in a near-to-far mode. Correspondingly, the calculation priority specifically refers to the calculation priority of the target grid size of the grids to be divided, namely, the closer the distance between the grids to be divided and the target object is, the higher the calculation priority of the target grid size is, so that the grid size is enabled to linearly increase on the basis of grid distribution parameters, and the effect that the grid distribution is enabled to be close to distant is achieved.

Based on the above, when performing grid division processing according to a preset non-uniform grid division policy, the reference grid size and the grid distribution parameters can be determined according to the preset non-uniform division policy; then, calculating the target grid size of the grids to be divided in the grid queue to be divided according to the reference grid size and the grid distribution parameters; in the calculation process, the grids to be divided in the grid queue to be divided are ordered according to the interval between the grids to be divided and the target grid size of the grids to be divided, and the calculation priority is determined according to the ordering result of the grids to be divided in the grid queue to be divided, so that the calculation sequence of the grid size corresponding to each grid to be divided is determined by the distance between the grids to be divided and the target object, and the grid size linearly grows according to the grid distribution parameters, thereby obtaining the near-dense and far-sparse grid size, and then the grid division processing is carried out on the grid graph to be divided according to the target grid size of the grids to be divided in the grid queue to be divided, so that the near-dense and far-sparse target grid graph of the grids can be obtained.

In addition, considering that the grid distribution parameters are the basis for controlling grid division and the linear distribution of the grid sizes needs to be met, so that the divided grid graph achieves the effect of near density and far thinning, the grid distribution parameters can be combined with the upper limit and the lower limit of the grid sizes to finish calculation, and the obtained grid distribution parameters can be finished in a linear growth trend when each grid size is calculated; in this embodiment, the specific implementation manner is as follows:

determining the reference grid size and the associated grid size according to a preset non-uniform division strategy; constructing the grid queues to be partitioned according to the grid graph to be partitioned, and determining the number of grids to be partitioned contained in the grid queues to be partitioned; and determining a grid size difference value based on the reference grid size and the associated grid size, and calculating the grid distribution parameter according to the grid size difference value and the number of grids to be divided.

Specifically, the associated grid size refers to a maximum value of the grid size, and in practical application, the reference grid size and the associated grid size may be set according to practical requirements, which is not limited in this embodiment. Correspondingly, the number of grids to be divided specifically refers to the number of grids to be constructed in the currently constructed grid map to be divided.

Based on the above, when determining the grid distribution parameters, the reference grid size and the associated grid size can be determined according to a preset non-uniform division strategy; thereafter, a grid queue to be partitioned can be constructed for the grid graph to be partitioned, and the number of grids to be partitioned contained in the grid queue to be partitioned is determined; based on the grid distribution parameters, the grid size difference is determined based on the reference grid size and the associated grid size, and the grid distribution parameters can be calculated according to the grid size difference and the number of grids to be divided, so that the grid distribution parameters are used for grid division processing.

In specific implementation, when calculating the grid size, the calculation may be performed with the lower boundary x_min=0 of the point cloud range, and the relationship between the grid size and the grid coordinates is as follows formula (1):

v _n ＝kn+b，k＝(v _max -v _min )/(x _max /v-1)，b＝-v _min ； (1)

wherein v represents the average grid size, v _max Representation ofMaximum grid size, v _min Representing the minimum grid size, x _max Indicating the up-limit of the point cloud, n indicates the nth grid, v _n Represents the size of the nth grid and k represents the linear growth slope.

Therefore, the size of each grid can be calculated by combining the relationship between the grid size and the grid coordinates, and the grid division is performed according to the calculated grid size, so that the near-dense far-sparse target grid graph can be obtained.

Along the above example, after obtaining the grid diagram to be divided including the unmanned distribution vehicle a, the calculation of the grid size may be performed according to the near-dense and far-sparse grid division manner, that is, by combining the above formula (1), it is determined that n grids are included in the grid diagram to be divided, and the n grids are ordered according to the distance from the unmanned distribution vehicle a in the grid diagram to be divided. Calculating the size of each grid from near to far according to the sorting result, determining the 1 st grid size as 0*0, the corresponding grid coordinate as (0, 0), the second grid size as 0.1 x 0.1, the corresponding grid coordinate as (0.1), the third grid size as 0.3 x 0.3, and the corresponding grid coordinate as (0.3 ) … …; after the grid sizes corresponding to the n grids are calculated, carrying out grid division processing on the grid graph to be divided according to the grid sizes, and obtaining a target grid graph shown in fig. 3 according to the division result; it should be noted that, the corresponding grid sizes after the mth grid in the target grid graph are the same.

In summary, the grid division processing is performed on the grid graph to be divided by adopting a near-dense and far-sparse grid division mode, so that the grid which is closer to the target object in the finally obtained target grid graph is denser and the grid which is farther to the target object is sparser, more computing resources can be distributed on the short-distance computing when the subsequent point cloud rasterization processing is performed, and fewer computing resources are adopted in a long distance, so that the resource utilization rate is effectively improved.

Step S206, collecting point cloud data aiming at the target object, and constructing a coordinate mapping relation between the point cloud data and grids in the target grid graph.

Specifically, after the near-dense and far-sparse target raster image is obtained, further, as the grids contained in the target raster image are changed, the position coordinates of each grid are changed, if the uniform raster image is still used for carrying out the point cloud rasterization processing, the point cloud data cannot be mapped to the real position, and therefore the driving safety of the target object is affected. Therefore, in order to ensure that the point cloud data can still be accurately mapped into the corresponding grid, after the point cloud data is acquired, the grid coordinates are repositioned for the non-uniform grid in the grid graph, and then a coordinate mapping relation is established between the point cloud data and the grid with the repositioned grid coordinates, so that the point cloud data can be accurately mapped into the target grid graph for subsequent processing according to the coordinate mapping relation.

The point cloud data specifically refers to a three-dimensional point data set of the surface of an actual road, which is acquired by a three-dimensional laser scanning instrument, and the three-dimensional points include, but are not limited to, three-dimensional points corresponding to any object on the road, such as vehicles, people, trees, roadblocks and the like; correspondingly, the coordinate mapping relationship is specifically a mapping relationship between point cloud coordinates in the point cloud data and grid coordinates of grids in the target grid graph, and each grid coordinate may correspond to one or more point cloud coordinates.

Further, when the coordinate mapping relation between the point cloud data and the grids in the target grid graph is constructed, the point cloud data contains a large number of point cloud coordinates, each grid in the target grid graph corresponds to different grid coordinates, and the result of the point cloud coordinate mapping to the grid meets the position relation under the real driving scene, so that the safe driving of the target object can be ensured, and the matching between the point cloud coordinates and the grid coordinates can be performed according to the point cloud rasterization strategy; in this embodiment, the specific implementation manner is as follows:

determining point cloud coordinates contained in the point cloud data and grid coordinates corresponding to grids in the target grid graph; and matching the point cloud coordinates with the grid coordinates according to a preset point cloud rasterization strategy, and constructing the coordinate mapping relation according to a matching result.

Specifically, the point cloud coordinates specifically refer to coordinates of three-dimensional points of the actual road surface acquired by the three-dimensional laser scanning instrument. Correspondingly, the grid coordinates specifically refer to coordinates corresponding to each grid in the target grid graph, and as each grid is a grid obtained by dividing according to a non-uniform dividing strategy and the grid sizes may be different, each grid coordinate can be determined according to the size of each grid, and the grid coordinates are two-dimensional coordinates; mapping the point cloud coordinates to the grid map is the process of mapping the three-dimensional point coordinates to the two-dimensional space. Correspondingly, the point cloud rasterization strategy specifically refers to a strategy for matching grid coordinates and point cloud coordinates, and the strategy can map three-dimensional points under real road conditions to positions in a two-dimensional space, so that the grid coordinates with a mapping relation with the point cloud coordinates are determined, the mapping relation before the two points cloud coordinates are formed, and after all the mapping relations are established, a coordinate mapping relation is obtained for use.

Based on the above, after the point cloud data of the associated target object is acquired, the point cloud coordinates contained in the point cloud data and the grid coordinates corresponding to the grids in the target grid graph can be determined; at this time, the point cloud coordinates and the grid coordinates can be matched according to a preset point cloud rasterization strategy, so that a coordinate mapping relation is constructed according to a matching result, and the subsequent mapping processing of the point cloud data according to the mapping relation is facilitated.

In practical application, when matching processing between the point cloud coordinates and the grid coordinates is performed, the following formula (2) may be adopted to complete matching:

wherein x represents the point cloud coordinates, and the explanation of other related parameters can be referred to the explanation of related parameters in the above formula (1), which is not repeated here. Through the formula (2), when the point cloud coordinates and the grid coordinates are matched, the coordinates can be substituted into the formula (2) to perform equation matching, and the coordinates forming a correct equation are the mapping relationship.

Along the above example, after the three-dimensional laser scanning instrument configured for the unmanned distribution vehicle A collects the point cloud data, matching between the point cloud coordinates and the grid coordinates in the point cloud data is carried out by combining the formula (2), and the point cloud coordinates x1 and the grid coordinates s1 are determined to be matched according to the matching result; the point cloud coordinate x2 is matched with the grid coordinate s2, the … … point cloud coordinate x is matched with the grid coordinate s, and after all matching relations are determined, the point cloud coordinate can be used for mapping the point cloud coordinate into a target grid graph later, so that the unmanned distribution vehicle A can safely and normally run.

In conclusion, the matching of the point cloud coordinates and the grid coordinates is performed by combining the point cloud rasterization strategy, so that the coordinate mapping relation between the point cloud and the grid can be accurately established, the point cloud rasterization processing is performed based on the relation, and the mapping accuracy can be ensured.

Step S208, mapping the point cloud data to the target raster pattern based on the coordinate mapping relation.

Specifically, after the establishment of the coordinate mapping relationship between the point cloud data and the grids in the target grid graph is completed, further, the point cloud data can be mapped into the target grid graph based on the coordinate mapping relationship, so that the target grid graph mapped with the point cloud data is obtained, and safe driving of the target object is realized.

Further, when mapping the point cloud data to the target grid graph, the point cloud coordinates need to be mapped to the corresponding grids in combination with the coordinate mapping relationship, so that the subsequent use can be facilitated, and in this embodiment, the specific implementation manner is as follows:

determining target grid coordinates of a corresponding grid of a target point Yun Zuobiao in the point cloud data in the target grid graph according to the coordinate mapping relation; and mapping the target point cloud coordinates to target grids corresponding to the target grid coordinates, and mapping the point cloud data to the target grid map.

Specifically, the target point cloud coordinate specifically refers to a point cloud coordinate which needs to be mapped at the current moment; correspondingly, the target grid coordinate specifically refers to a grid coordinate with a mapping relation with the cloud coordinate of the target point. Based on the above, when mapping processing is performed, according to the coordinate mapping relation, determining the target grid coordinates of the corresponding grid of the target point cloud coordinates in the target grid graph in the target point cloud data; at this time, the cloud coordinates of the target point can be mapped to the target grid corresponding to the target grid coordinates, and the like until all the cloud coordinates of the target point are mapped.

Along the above example, after determining the mapping relationship between the point cloud coordinates and the grid coordinates, the grid v1 in the target grid diagram corresponding to the grid coordinate v1 may be determined according to the coordinate mapping relationship, the grid v2 in the target grid diagram corresponding to the grid v2 … … grid coordinate v in the target grid diagram corresponds to the grid v in the target grid diagram, at this time, the grid diagram as shown in fig. 4 may be obtained according to the mapping result, and then the running calculation processing of the unmanned distribution vehicle a may be performed by using the grid diagram to control the safe and stable running of the unmanned distribution vehicle a.

In conclusion, the target grids are positioned by combining the coordinate mapping relation and the point cloud rasterization processing is performed, so that the accuracy of the point cloud rasterization can be ensured, and the safety and reliability are realized in the running of the target object.

In addition, in the application stage, after the quantity point cloud data is mapped into the target raster pattern, in order to drive the target object, automatic driving can be carried out according to the actual road condition, and detection of the concerned object can be carried out; in this embodiment, the specific implementation manner is as follows:

generating an object detection diagram according to the mapping result; performing object detection of interest on the object detection graph, and creating an object adjustment task according to a detection result; and driving the target object to execute the object adjustment task.

Specifically, the object detection map specifically refers to a two-dimensional map generated according to a target raster map of the mapping point cloud data, and is used for detecting an object of interest in the object detection map, so as to plan a driving route of the target object. Accordingly, the object of interest specifically refers to an object of interest in the current driving scene of the target object, which may affect driving, such as vehicles, pedestrians, road blocks, and the like. Correspondingly, the object adjustment task specifically refers to a task of driving the target object to perform route adjustment, including but not limited to driving the target object at a constant speed, accelerating, changing lanes, braking, and the like.

Based on this, after generating the object detection map from the mapping result; further, in order to enable the target object to drive normally, object detection is performed with respect to the object detection diagram, and an object adjustment task is created according to the detection result; thus, the target object is driven to execute the object adjustment task, and smooth running on the road is realized.

In summary, by performing object-of-interest detection in combination with the object detection graph, an object adjustment task satisfying the current driving scene can be created, and the target object is driven to execute the task, so that driving safety can be ensured.

Furthermore, when performing object detection and creating an object adjustment task, the method may be implemented in combination with a neural network model, and in this embodiment, the specific implementation manner is as follows:

inputting the object detection graph into a target detection model for processing, and obtaining the object detection graph containing the object of interest according to a processing result; cutting an object detection diagram containing the object of interest according to the object position information corresponding to the target object to obtain an object adjustment diagram; and creating an object adjustment task associated with the target object according to the object adjustment graph.

Specifically, the object detection model specifically refers to a model with object detection capability, and can identify and frame the object of interest in the object detection graph. Correspondingly, the object position information specifically refers to the position of the target object in the object detection diagram; correspondingly, the object adjustment map specifically refers to a two-dimensional map obtained by clipping an object detection map containing the object of interest. Considering that the object detection diagram may contain some contents which do not affect driving, in order to reduce resource consumption, the area affecting driving may be cut off according to the position of the target object, so as to obtain the object adjustment diagram.

Based on the above, the object detection map may be input to the target detection model for processing, and the object detection map including the object of interest may be obtained according to the processing result; then, in order to reduce the consumption of computing resources, the object detection diagram containing the object of interest can be cut according to the object position information corresponding to the target object, so that an object adjustment diagram is obtained according to a cutting result; and then, creating an object adjustment task of the associated target object according to the object adjustment graph.

According to the above example, after the target raster pattern of the mapping point cloud data is obtained, the target raster pattern can be input into a detection model for processing, so that an object affecting the running of the unmanned distribution vehicle A in an image is identified, an electric vehicle is determined to be located at the distance L in front of the unmanned distribution vehicle A according to the identification result, the running route of the electric vehicle is overlapped with the running route of the unmanned distribution vehicle A, traffic accidents are avoided, the unmanned distribution vehicle A can be controlled to change lanes to the left side, and the electric vehicle can safely run on a road.

The application of the point cloud rasterization method provided in the present specification in an autopilot scenario is taken as an example in the following description with reference to fig. 5. Fig. 5 shows a process flow chart of a point cloud rasterizing method according to an embodiment of the present disclosure, which specifically includes the following steps.

Step S502, loading a target virtual map according to the position information of the target vehicle, and constructing an initial raster image according to the target virtual map.

Step S504, updating the initial raster pattern based on the position information and the attribute information of the target vehicle, and generating the raster pattern to be divided containing the target vehicle according to the updating result.

Step S506, determining the reference grid size and the grid distribution parameters according to a preset non-uniform division strategy.

Step S508, calculating the target grid size of the grids to be divided in the grid to be divided queue according to the reference grid size and the grid distribution parameters.

Step S510, according to the target grid size of the grids to be divided in the grid queue to be divided, performing grid division processing on the grids to be divided to obtain target grids.

The grids to be divided in the grid queue to be divided are ordered according to the interval between the grids to be divided and the target vehicles, the calculation priority of the target grid size of the grids to be divided is determined according to the ordering result of the grids to be divided in the grid queue to be divided, and the size of the grids in the target grid graph is in direct proportion to the interval between the grids and the target vehicles;

specifically, the determining of the grid distribution parameters includes: determining a reference grid size and an associated grid size according to a preset non-uniform division strategy; constructing a grid queue to be partitioned according to the grid graph to be partitioned, and determining the number of grids to be partitioned contained in the grid queue to be partitioned; and determining a grid size difference value based on the reference grid size and the associated grid size, and calculating grid distribution parameters according to the grid size difference value and the number of grids to be divided.

In step S512, the point cloud coordinates included in the point cloud data and the grid coordinates corresponding to the grids in the target grid graph are determined.

Step S514, matching the point cloud coordinates and the grid coordinates according to a preset point cloud rasterization strategy, and constructing a coordinate mapping relation according to a matching result.

Step S516, according to the coordinate mapping relation, determining the target grid coordinates of the corresponding grids of the target point cloud coordinates in the target grid graph in the target point cloud data.

In step S518, the target point cloud coordinates are mapped to the target grids corresponding to the target grid coordinates, and the point cloud data are mapped to the target grid map.

Step S520, a vehicle detection map is generated according to the mapping result, the vehicle detection map is input to the target detection model for processing, and a vehicle detection map including the vehicle of interest is obtained according to the processing result.

Step S522, clipping the vehicle detection map including the vehicle of interest according to the vehicle position information corresponding to the target vehicle, and obtaining a vehicle adjustment map.

Step S524, creating a vehicle adjustment task associated with the target vehicle according to the vehicle adjustment map, and driving the target vehicle to execute the vehicle adjustment task.

Corresponding to the above method embodiments, another embodiment of a method for rasterizing a point cloud is further provided in the present specification, and fig. 6 is a schematic diagram illustrating another method for rasterizing a point cloud provided in one embodiment of the present specification. As shown in fig. 6, the method is applied to a vehicle control end and comprises the following steps of

Step S602, obtaining a grid diagram to be divided containing a target vehicle;

step S604, carrying out grid division processing on the grid graph to be divided according to a preset non-uniform division strategy to obtain a target grid graph, wherein the size of a grid in the target grid graph is in direct proportion to the interval of the grid from the target vehicle;

step S606, collecting point cloud data aiming at the target vehicle, and constructing a coordinate mapping relation between the point cloud data and grids in the target grid graph;

step S608, mapping the point cloud data to the target raster pattern based on the coordinate mapping relationship, wherein the target raster pattern mapped with the point cloud data is used for automatic driving of the target vehicle.

The other point cloud rasterization method provided in this embodiment is similar to the point cloud rasterization method in the above embodiment, and the same or corresponding description contents can be referred to the above embodiment, which is not repeated here.

Corresponding to the method embodiment, the present disclosure further provides an embodiment of a point cloud rasterizing device, and fig. 7 shows a schematic structural diagram of the point cloud rasterizing device according to one embodiment of the present disclosure. As shown in fig. 7, the apparatus includes:

an obtaining module 702 configured to obtain a grid pattern to be divided including a target object;

the dividing module 704 is configured to perform grid division processing on the grid graph to be divided according to a preset non-uniform division strategy to obtain a target grid graph, wherein the size of a grid in the target grid graph is in direct proportion to the interval between the grid and the target object;

a construction module 706, configured to collect point cloud data for the target object, and construct a coordinate mapping relationship between the point cloud data and a grid in the target grid graph;

a mapping module 708 is configured to map the point cloud data to the target raster pattern based on the coordinate mapping relationship.

In an alternative embodiment, the obtaining module 702 is further configured to:

In an alternative embodiment, the partitioning module 704 is further configured to:

In an alternative embodiment, the building block 706 is further configured to:

In an alternative embodiment, the apparatus further comprises:

the detection module is configured to generate an object detection diagram according to the mapping result;

performing object detection of interest on the object detection graph, and creating an object adjustment task according to a detection result;

and driving the target object to execute the object adjustment task.

In an alternative embodiment, the detection module is further configured to:

In an alternative embodiment, the mapping module 708 is further configured to:

In the point cloud rasterizing device provided by the specification, in order to reduce the consumption of calculation performance while improving the resolution of a region of interest, a near-dense and far-sparse non-uniform rasterizing scheme is adopted for raster division. Namely: the method comprises the steps of firstly obtaining a grid diagram to be divided containing a target object, and then carrying out grid division processing on the grid diagram to be divided according to a preset non-uniform division strategy in order to balance the resolution and the consumption of computing resources, so that the resolution corresponding to a region of interest with a relatively close distance to the target object is improved, and meanwhile, less computing resources are used. In the grid dividing process, the grid size in the grid chart is directly proportional to the interval between the grids and the target object, that is, the grid size closer to the target object is smaller, the grid size farther from the target object is larger, the grid dividing process is completed, the near-density and far-thinning process is achieved, the divided non-uniform grids achieve higher processing resolution closer to the target object, more resource consumption, lower processing resolution and less resource consumption in the subsequent process, and therefore the resource consumption is effectively controlled. On the basis, after the point cloud data are acquired, the coordinate position of each grid is not fixed in consideration of the fact that the grids are non-uniform grids, so that the coordinate mapping relation between the point cloud data and the grids in the target grid graph can be constructed firstly, and the mapping relation between the point cloud coordinates in the point cloud data and the grids in the target grid graph can be determined; based on the mapping relation, the point cloud data can be mapped into the target raster pattern, the mapping result is consistent with the position relation in the real scene, and the processing of the downstream task can be realized by realizing the target raster pattern for mapping the point cloud data, so that the aims of balancing the resolution and the resource consumption in the point cloud rasterization processing are fulfilled.

The above is a schematic scheme of a point cloud rasterizing apparatus of this embodiment. It should be noted that, the technical solution of the point cloud rasterizing device and the technical solution of the point cloud rasterizing method belong to the same concept, and details of the technical solution of the point cloud rasterizing device, which are not described in detail, can be referred to the description of the technical solution of the point cloud rasterizing method.

Corresponding to the method embodiment, the present disclosure further provides another embodiment of the point cloud rasterizing device, and fig. 8 shows a schematic structural diagram of another point cloud rasterizing device provided in one embodiment of the present disclosure. As shown in fig. 8, the device is applied to a vehicle control end, and comprises:

an acquire raster pattern module 802 configured to acquire a raster pattern to be divided including a target vehicle;

the grid pattern dividing module 804 is configured to perform grid division processing on the grid pattern to be divided according to a preset non-uniform division policy to obtain a target grid pattern, wherein the size of a grid in the target grid pattern is in direct proportion to the interval between the grid and the target vehicle;

a build mapping relationship module 806 configured to collect point cloud data for the target vehicle, and build a coordinate mapping relationship between the point cloud data and a grid in the target grid graph;

And a map point cloud data module 808 configured to map the point cloud data to the target raster pattern based on the coordinate mapping relationship, wherein the target raster pattern to which the point cloud data is mapped is used for automatic driving of the target vehicle.

It should be noted that, according to the types of the automatic driving vehicles, the algorithm modules (the functional modules) mentioned in the above embodiments may also be different. For example, different algorithm modules may be involved for logistics vehicles, public service vehicles, medical service vehicles, terminal service vehicles. The algorithm module is illustrated below for each of these four autonomous vehicles:

the logistics vehicles refer to vehicles used in logistics scenes, and can be logistics vehicles with automatic sorting functions, logistics vehicles with refrigerating and heat-preserving functions and logistics vehicles with measuring functions. These logistics vehicles may involve different algorithm modules.

For example, for a logistics vehicle, an automated sorting device may be provided which can automatically pick up and transport, sort and store goods after the logistics vehicle arrives at the destination. This involves an algorithm module for sorting of goods, which mainly implements logic control of goods taking out, handling, sorting and storing.

For another example, for a cold chain logistics scene, the logistics vehicle can be further provided with a refrigeration and heat preservation device, and the refrigeration and heat preservation device can realize refrigeration or heat preservation of transported fruits, vegetables, aquatic products, frozen foods and other perishable foods, so that the fruits, vegetables, aquatic products, frozen foods and other perishable foods are in a proper temperature environment, and the problem of long-distance transportation of perishable foods is solved. The algorithm module is mainly used for dynamically and adaptively calculating proper temperature of cold food or heat preservation according to information such as food (or article) properties, perishability, transportation time, current seasons, weather and the like, and automatically adjusting the cold food or heat preservation device according to the proper temperature, so that transportation personnel do not need to manually adjust the temperature when different foods or articles are transported by a vehicle, the transportation personnel are liberated from complicated temperature regulation and control, and the efficiency of cold food or heat preservation transportation is improved.

For example, in most logistics scenes, the charge is carried out according to the volume and/or weight of the packages, the number of the logistics packages is very large, and the volume and/or weight of the packages are simply measured by an express delivery person, so that the efficiency is very low, and the labor cost is high. Therefore, in some logistics vehicles, a measuring device is additionally arranged, so that the volume and/or the weight of the logistics package can be automatically measured, and the cost of the logistics package can be calculated. This involves an algorithm module for logistic parcel measurement which is primarily used to identify the type of logistic parcel, determine the way in which the logistic parcel is measured, such as whether a volumetric measurement or a weight measurement is made or a combination of volumetric and weight measurements are made simultaneously, and can perform volumetric and/or weight measurements based on the determined way of measurement, and perform cost calculations based on the measurement results.

The public service vehicle refers to a vehicle for providing a certain public service, and can be, for example, a fire truck, a deicing vehicle, a sprinkler, a snow plow, a garbage disposal vehicle, a traffic guidance vehicle and the like. These public service vehicles may involve different algorithm modules.

For example, for an automatically driven fire engine, the main task is to perform a reasonable fire extinguishing task for a fire scene, which involves an algorithm module for the fire extinguishing task, and the algorithm module at least needs to implement logic of fire condition identification, fire extinguishing scheme planning, automatic control of a fire extinguishing device and the like.

For another example, for deicing vehicles, the main task is to remove ice and snow on the road surface, which involves an algorithm module for deicing that at least needs to implement logic for identifying ice and snow conditions on the road surface, making deicing schemes based on the ice and snow conditions, such as which road segments need to be defrosted, which road segments need not be defrosted, whether salt spraying mode, salt spraying gram number, etc. are used, and automatic control of the deicing device in case of determining the deicing scheme.

The medical service vehicle is an automatic driving vehicle capable of providing one or more medical services, and the vehicle can provide medical services such as disinfection, temperature measurement, medicine preparation, isolation and the like, and the medical service vehicle relates to algorithm modules for providing various self-service medical services, wherein the algorithm modules mainly realize the identification of disinfection requirements and the control of disinfection devices so as to enable the disinfection devices to disinfect patients or identify the positions of the patients, control the temperature measurement devices to automatically measure the temperature of the patients at the positions of the forehead and the like of the patients, or realize the judgment of symptoms, give medicine according to the judgment result and need to realize the identification of medicines/medicine containers, control the medicine taking mechanical arm so as to enable the medicine taking mechanical arm to take medicines for the patients according to the medicine prescription, and the like.

The terminal service vehicle refers to a self-service type automatic driving vehicle capable of replacing some terminal equipment to provide certain convenience services for users, for example, the vehicle can provide printing, attendance checking, scanning, unlocking, payment, retail and other services for the users.

For example, in some application scenarios, users often need to go to a particular location to print or scan a document, which is time consuming and laborious. Therefore, there is a terminal service vehicle capable of providing a printing/scanning service for a user, the service vehicles can be interconnected with a user terminal device, the user sends a printing command through the terminal device, the service vehicle responds to the printing command, automatically prints a document required by the user and can automatically send the printed document to a user position, the user does not need to go to a printer for queuing, and the printing efficiency can be greatly improved. Or, the user can respond to the scanning instruction sent by the terminal equipment and move to the user position, and the user can finish scanning on the scanning tool of the service vehicle for placing the document to be scanned, so that queuing at a printer/scanner is not needed, and time and labor are saved. This involves an algorithm module providing print/scan services that at least needs to identify interconnections with the user terminal device, responses to print/scan instructions, positioning of user location, travel control, etc.

For another example, as new retail projects develop, more and more electronic commerce uses self-service vending machines to sell goods to various office buildings and public areas, but the self-service vending machines are placed at fixed positions and are not movable, and users need to go to the self-service vending machines before they can purchase the required goods, so that convenience is still poor. The self-service driving vehicles capable of providing retail services are arranged, the service vehicles can bear goods to automatically move, corresponding self-service shopping APP or shopping portals can be provided, a user can place an order to the self-service driving vehicles providing retail services through the APP or shopping portals by means of terminals such as mobile phones, the order comprises names, quantity and user positions of goods to be purchased, after receiving an order placing request, the vehicles can determine whether the current remaining goods have the goods purchased by the user and whether the quantity is enough, and under the condition that the goods purchased by the user are determined to be enough, the goods can be carried to the user positions automatically, and the goods are provided for the user, so that the convenience of shopping of the user is further improved, the user time is saved, and the user can use the time for more important things. This involves the algorithm modules providing retail services that implement mainly logic for responding to user order requests, order processing, merchandise information maintenance, user location positioning, payment management, etc.

The above is another exemplary embodiment of the point cloud rasterizing apparatus of the present embodiment. It should be noted that, the technical solution of the point cloud rasterizing device and the technical solution of the point cloud rasterizing method belong to the same concept, and details of the technical solution of the point cloud rasterizing device, which are not described in detail, can be referred to the description of the technical solution of the point cloud rasterizing method.

Fig. 9 illustrates a block diagram of a computing device 900 provided in accordance with one embodiment of the present specification. The components of computing device 900 include, but are not limited to, memory 910 and processor 920. Processor 920 is coupled to memory 910 via bus 930 with database 950 configured to hold data.

Computing device 900 also includes an access device 940, access device 940 enabling computing device 900 to communicate via one or more networks 960. Examples of such networks include public switched telephone networks (PSTN, public Switched Telephone Network), local area networks (LAN, local Area Network), wide area networks (WAN, wide Area Network), personal area networks (PAN, personal Area Network), or combinations of communication networks such as the internet. Access device 940 may include one or more of any type of network interface, wired or wireless, such as a network interface card (NIC, network interface controller), such as an IEEE802.11 wireless local area network (WLAN, wireless Local Area Network) wireless interface, a worldwide interoperability for microwave access (Wi-MAX, worldwide Interoperability for Microwave Access) interface, an ethernet interface, a universal serial bus (USB, universal Serial Bus) interface, a cellular network interface, a bluetooth interface, near field communication (NFC, near Field Communication).

In one embodiment of the present description, the above-described components of computing device 900 and other components not shown in FIG. 9 may also be connected to each other, for example, by a bus. It should be understood that the block diagram of the computing device illustrated in FIG. 9 is for exemplary purposes only and is not intended to limit the scope of the present description. Those skilled in the art may add or replace other components as desired.

Computing device 900 may be any type of stationary or mobile computing device including a mobile computer or mobile computing device (e.g., tablet, personal digital assistant, laptop, notebook, netbook, etc.), mobile phone (e.g., smart phone), wearable computing device (e.g., smart watch, smart glasses, etc.), or other type of mobile device, or a stationary computing device such as a desktop computer or personal computer (PC, personal Computer). Computing device 900 may also be a mobile or stationary server.

The processor 920 is configured to execute computer-executable instructions that, when executed by the processor, implement the steps of the point cloud rasterization method described above.

The foregoing is a schematic illustration of a computing device of this embodiment. It should be noted that, the technical solution of the computing device and the technical solution of the above-mentioned point cloud rasterization method belong to the same concept, and details of the technical solution of the computing device, which are not described in detail, can be referred to the description of the technical solution of the above-mentioned point cloud rasterization method.

An embodiment of the present disclosure also provides a computer-readable storage medium storing computer-executable instructions that, when executed by a processor, implement the steps of the point cloud rasterizing method described above.

The above is an exemplary version of a computer-readable storage medium of the present embodiment. It should be noted that, the technical solution of the storage medium and the technical solution of the above-mentioned point cloud rasterization method belong to the same concept, and details of the technical solution of the storage medium which are not described in detail can be referred to the description of the technical solution of the above-mentioned point cloud rasterization method.

An embodiment of the present disclosure further provides a computer program, where the computer program, when executed in a computer, causes the computer to perform the steps of the point cloud rasterizing method described above.

The above is an exemplary version of a computer program of the present embodiment. It should be noted that, the technical solution of the computer program and the technical solution of the point cloud rasterization method belong to the same concept, and details of the technical solution of the computer program, which are not described in detail, can be referred to the description of the technical solution of the point cloud rasterization method.

The foregoing describes specific embodiments of the present disclosure. Other embodiments are within the scope of the following claims. In some cases, the actions or steps recited in the claims can be performed in a different order than in the embodiments and still achieve desirable results. In addition, the processes depicted in the accompanying figures do not necessarily require the particular order shown, or sequential order, to achieve desirable results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

The computer instructions include computer program code that may be in source code form, object code form, executable file or some intermediate form, etc. The computer readable medium may include: any entity or device capable of carrying the computer program code, a recording medium, a U disk, a removable hard disk, a magnetic disk, an optical disk, a computer Memory, a Read-Only Memory (ROM), a random access Memory (RAM, random Access Memory), an electrical carrier signal, a telecommunications signal, a software distribution medium, and so forth. It should be noted that the content of the computer readable medium can be increased or decreased appropriately according to the requirements of the patent practice, for example, in some areas, according to the patent practice, the computer readable medium does not include an electric carrier signal and a telecommunication signal.

It should be noted that, for simplicity of description, the foregoing method embodiments are all expressed as a series of combinations of actions, but it should be understood by those skilled in the art that the embodiments are not limited by the order of actions described, as some steps may be performed in other order or simultaneously according to the embodiments of the present disclosure. Further, those skilled in the art will appreciate that the embodiments described in the specification are all preferred embodiments, and that the acts and modules referred to are not necessarily all required for the embodiments described in the specification.

In the foregoing embodiments, the descriptions of the embodiments are emphasized, and for parts of one embodiment that are not described in detail, reference may be made to the related descriptions of other embodiments.

The preferred embodiments of the present specification disclosed above are merely used to help clarify the present specification. Alternative embodiments are not intended to be exhaustive or to limit the invention to the precise form disclosed. Obviously, many modifications and variations are possible in light of the teaching of the embodiments. The embodiments were chosen and described in order to best explain the principles of the embodiments and the practical application, to thereby enable others skilled in the art to best understand and utilize the invention. This specification is to be limited only by the claims and the full scope and equivalents thereof.

Claims

1. A method of point cloud rasterization, comprising:

obtaining a grid diagram to be divided containing a target object;

2. The method for rasterizing a point cloud according to claim 1, wherein the obtaining a raster pattern to be divided including a target object includes:

loading a target virtual map according to the position information of the target object, and constructing an initial grid map according to the target virtual map;

updating the initial raster image based on the position information and the attribute information of the target object, and generating a raster image to be divided containing the target object according to an updating result.

3. The method of rasterizing a point cloud according to claim 1, wherein the performing raster division processing on the raster image to be divided according to a preset non-uniform division policy, to obtain a target raster image, includes:

Determining a reference grid size and grid distribution parameters according to a preset non-uniform division strategy;

calculating the target grid size of the grids to be divided in the grid queue to be divided according to the reference grid size and the grid distribution parameters;

according to the target grid size of grids to be divided in the grid queue to be divided, performing grid division processing on the grid graph to be divided to obtain the target grid graph;

the grids to be divided in the grid queue to be divided are ordered according to the interval with the target object, and the calculation priority of the target grid size of the grids to be divided is determined according to the ordering result of the grids to be divided in the grid queue to be divided.

4. A method of point cloud rasterizing as claimed in claim 3, the determination of the grid distribution parameters comprising:

determining the reference grid size and the associated grid size according to a preset non-uniform division strategy;

constructing the grid queues to be partitioned according to the grid graph to be partitioned, and determining the number of grids to be partitioned contained in the grid queues to be partitioned;

and determining a grid size difference value based on the reference grid size and the associated grid size, and calculating the grid distribution parameter according to the grid size difference value and the number of grids to be divided.

5. The method of point cloud rasterizing of claim 1, the constructing a coordinate mapping relationship between the point cloud data and a grid in the target raster image, comprising:

determining point cloud coordinates contained in the point cloud data and grid coordinates corresponding to grids in the target grid graph;

and matching the point cloud coordinates with the grid coordinates according to a preset point cloud rasterization strategy, and constructing the coordinate mapping relation according to a matching result.

6. The method of point cloud rasterizing according to claim 1, wherein after the step of mapping the point cloud data to the target raster pattern is performed based on the coordinate mapping relationship, further comprising:

generating an object detection diagram according to the mapping result;

and driving the target object to execute the object adjustment task.

7. The method of point cloud rasterizing according to claim 6, wherein the object detection map performs object of interest detection, and creates an object adjustment task according to a detection result, including:

inputting the object detection graph into a target detection model for processing, and obtaining the object detection graph containing the object of interest according to a processing result;

Cutting an object detection diagram containing the object of interest according to the object position information corresponding to the target object to obtain an object adjustment diagram;

and creating an object adjustment task associated with the target object according to the object adjustment graph.

8. The method of rasterizing a point cloud of any one of claims 1-7, the mapping the point cloud data to the target raster pattern based on the coordinate mapping relationship, comprising:

determining target grid coordinates of a corresponding grid of a target point Yun Zuobiao in the point cloud data in the target grid graph according to the coordinate mapping relation;

and mapping the target point cloud coordinates to target grids corresponding to the target grid coordinates, and mapping the point cloud data to the target grid map.

9. A point cloud rasterization method is applied to a vehicle control end and comprises the following steps:

acquiring a grid diagram to be divided containing a target vehicle;

10. A computing device, comprising:

a memory and a processor;

the memory is configured to store computer executable instructions, the processor being configured to execute the computer executable instructions, which when executed by the processor, implement the steps of the method of any one of claims 1 to 9.

11. A computer readable storage medium storing computer executable instructions which when executed by a processor implement the steps of the method of any one of claims 1 to 9.