WO2022099511A1 - 基于点云数据的地面分割方法、装置和计算机设备 - Google Patents

基于点云数据的地面分割方法、装置和计算机设备 Download PDF

Info

Publication number
WO2022099511A1
WO2022099511A1 PCT/CN2020/128139 CN2020128139W WO2022099511A1 WO 2022099511 A1 WO2022099511 A1 WO 2022099511A1 CN 2020128139 W CN2020128139 W CN 2020128139W WO 2022099511 A1 WO2022099511 A1 WO 2022099511A1
Authority
WO
WIPO (PCT)
Prior art keywords
cloud data
point cloud
point
ground segmentation
target
Prior art date
Application number
PCT/CN2020/128139
Other languages
English (en)
French (fr)
Inventor
吴伟
Original Assignee
深圳元戎启行科技有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 深圳元戎启行科技有限公司 filed Critical 深圳元戎启行科技有限公司
Priority to PCT/CN2020/128139 priority Critical patent/WO2022099511A1/zh
Priority to CN202080093095.XA priority patent/CN114981840A/zh
Publication of WO2022099511A1 publication Critical patent/WO2022099511A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/10—Segmentation; Edge detection
    • G06T7/11—Region-based segmentation

Definitions

  • the present application relates to a ground segmentation method, device, computer equipment and storage medium based on point cloud data.
  • Lidar sensors can provide real-time and accurate three-dimensional scene information, with a large ranging range and high accuracy, and the measured values are not affected by ambient light.
  • the information collected by lidar sensors is usually presented in the form of point clouds.
  • ground segmentation of point cloud data is mainly performed by ground fitting or methods based on depth images and geometric relationships.
  • these methods are only suitable for ground segmentation of point cloud data in specific scenarios.
  • ground fitting methods are only suitable for flat roads.
  • Methods based on depth images and geometric relationships need to ensure that the installation of lidar sensors is in a horizontal state. , when the scene changes, if the ground segmentation of the point cloud data is performed in the traditional way, the accuracy of the ground segmentation of the point cloud data will be low.
  • a method, apparatus, computer device and storage medium for ground segmentation based on point cloud data are provided.
  • a ground segmentation method based on point cloud data comprising:
  • Extract the point cloud data in the three-dimensional bounding box select the target point cloud data in the extracted point cloud data, and perform plane fitting on the target point cloud data corresponding to the three-dimensional bounding box to obtain a fitting plane;
  • the ground segmentation result is updated according to the location area and the fitting plane, and the target ground segmentation result corresponding to the point cloud data is obtained.
  • a ground segmentation device based on point cloud data comprising:
  • the ground segmentation module is used to perform ground segmentation on the point cloud data to obtain the ground segmentation result
  • the target detection module is used to perform target detection on the point cloud data, and obtain the three-dimensional bounding box of the target object corresponding to the point cloud data;
  • the plane fitting module is used to extract the point cloud data in the three-dimensional bounding box, select the target point cloud data from the extracted point cloud data, and perform plane fitting on the target point cloud data corresponding to the three-dimensional bounding box to obtain the fitted plane;
  • an extraction module for extracting the location area corresponding to each target object according to the three-dimensional bounding box
  • the update module is used to update the ground segmentation result according to the location area and the fitting plane, and obtain the target ground segmentation result corresponding to the point cloud data.
  • a computer device includes a memory and one or more processors, the memory stores computer-readable instructions, and when the computer-readable instructions are executed by the processor, causes the one or more processors to perform the following steps:
  • Extract the point cloud data in the three-dimensional bounding box select the target point cloud data in the extracted point cloud data, and perform plane fitting on the target point cloud data corresponding to the three-dimensional bounding box to obtain a fitting plane;
  • the ground segmentation result is updated according to the location area and the fitting plane, and the target ground segmentation result corresponding to the point cloud data is obtained.
  • One or more non-volatile computer-readable storage media storing computer-readable instructions that, when executed by one or more processors, cause the one or more processors to perform the following steps:
  • Extract the point cloud data in the three-dimensional bounding box select the target point cloud data in the extracted point cloud data, and perform plane fitting on the target point cloud data corresponding to the three-dimensional bounding box to obtain a fitting plane;
  • the ground segmentation result is updated according to the location area and the fitting plane, and the target ground segmentation result corresponding to the point cloud data is obtained.
  • FIG. 1 is an application scenario diagram of a ground segmentation method based on point cloud data in one or more embodiments.
  • FIG. 2 is a schematic flowchart of a ground segmentation method based on point cloud data in one or more embodiments.
  • FIG. 3 is a schematic flowchart of a step of selecting target point cloud data from the extracted point cloud data in one or more embodiments.
  • Fig. 4 is a schematic flowchart of the steps of updating the ground segmentation result according to the location area and the fitting plane to obtain the target ground segmentation result in one or more embodiments.
  • FIG. 5 is a block diagram of an apparatus for ground segmentation based on point cloud data in one or more embodiments.
  • FIG. 6 is a block diagram of a computer device in one or more embodiments.
  • the ground segmentation method based on point cloud data provided in this application can be applied to the unmanned driving application scenario shown in FIG. 1 .
  • An on-board sensor 102 and on-board computer equipment 104 are pre-installed in the driverless vehicle.
  • the in-vehicle computer equipment may be simply referred to as computer equipment.
  • the vehicle-mounted sensor 102 communicates with the computer device 104 , and in the process of unmanned driving, the vehicle-mounted sensor 102 transmits the collected point cloud data to the computer device 104 .
  • the computer device 104 performs ground segmentation on the point cloud data to obtain a ground segmentation result.
  • the computer device 104 may also perform target detection on the point cloud data to obtain a three-dimensional bounding box of the target object corresponding to the point cloud data.
  • the computer device 104 extracts the point cloud data in the three-dimensional bounding box, selects target point cloud data from the extracted point cloud data, and performs plane fitting on the target point cloud data corresponding to the three-dimensional bounding box to obtain a fitted plane.
  • the computer device 104 extracts the position area corresponding to each target object according to the three-dimensional bounding box, so as to update the ground segmentation result according to the position area and the fitting plane, and obtain the target ground segmentation result corresponding to the point cloud data.
  • the onboard sensor 102 may be a lidar.
  • a ground segmentation method based on point cloud data is provided, and the method is applied to the computer equipment in FIG. 1 as an example to illustrate, including the following steps:
  • Step 202 acquiring point cloud data.
  • the point cloud data can be the data recorded by the vehicle-mounted sensor in the form of point cloud by scanning the surrounding environment information.
  • the point cloud data may specifically include three-dimensional coordinates (x, y, z) of each point, laser reflection intensity (Intensity), color information (RGB), and the like.
  • Three-dimensional coordinates are used to represent the position information of the target object surface in the surrounding environment.
  • the three-dimensional coordinates may be the coordinates of the point in the Cartesian coordinate system, and specifically include the horizontal, vertical, and vertical coordinates of the point in the Cartesian coordinate system.
  • the Cartesian coordinate system is a three-dimensional space coordinate system established with the vehicle-mounted sensor as the origin.
  • the three-dimensional space coordinate system includes a horizontal axis (x axis), a vertical axis (y axis) and a vertical axis (z axis).
  • the three-dimensional space coordinate system established with the vehicle sensor as the origin satisfies the right-hand rule.
  • the x-axis coordinate represents the longitudinal distance of the scanned target object surface relative to the on-board sensor
  • the y-axis coordinate represents the lateral offset of the scanned target object surface relative to the on-board sensor
  • the z-axis coordinate represents the scanned target.
  • the height of the object's surface relative to the onboard sensors is the height of the object's surface relative to the onboard sensors.
  • the current environment can be scanned by the on-board sensors installed on the vehicle to obtain the corresponding point cloud data, and the collected point cloud data can be transmitted to the computer equipment.
  • the onboard sensor could be a lidar.
  • Step 204 Perform ground segmentation on the point cloud data to obtain a ground segmentation result.
  • Ground segmentation refers to identifying ground points and non-ground points in point cloud data. Since the point cloud data includes a large amount of ground point cloud data, and the amount of useful information contained in the point cloud data can be used to complete tasks such as high-precision map reconstruction, obstacle detection and tracking, etc., it is necessary to analyze the point cloud data. Ground segmentation, so as to complete the target task according to the remaining point cloud data after ground point removal.
  • the computer equipment can perform ground segmentation on the point cloud data, thereby identifying the category of each point in the point cloud data, and obtaining corresponding ground segmentation results.
  • the categories specifically include ground points and non-ground points.
  • the ground segmentation method may be to first divide the point cloud area where the point cloud data is located into multiple sub-areas.
  • the point cloud area refers to the three-dimensional data space where the point cloud data is located.
  • the division method may be grid division of the point cloud area, that is, division of the horizontal plane formed by the point cloud area in the x-axis direction and the y-axis direction.
  • the grid division method may be equal division or random division.
  • the point cloud area can be equally divided into 10*10 horizontal grids.
  • the computer device may use the least squares method to estimate the corresponding ground according to the preset plane equation, so as to obtain the corresponding ground of each sub-region.
  • the preset plane equation may be a ternary linear equation.
  • the ground corresponding to each sub-region is embodied in the form of a ternary linear equation.
  • the computer equipment traverses the coordinates of the points in the corresponding sub-regions in the equation corresponding to the ground, calculates the distance between each point and the corresponding ground, and determines the point as a ground point when the distance is less than a threshold. When the distance is greater than or equal to the threshold, the point is determined as a non-ground point. Threshold refers to the distance threshold used to judge whether the point is a ground point.
  • the computer device may also adopt any of the point cloud ground segmentation methods that do not require standard ground data and training models, such as a method based on depth images and geometric relationships, a normal vector method, an absolute height method, and an average height method.
  • a ground segmentation for point cloud data may be adopted.
  • Step 206 Perform target detection on the point cloud data to obtain a three-dimensional bounding box of the target object corresponding to the point cloud data.
  • Target detection refers to identifying the target object category corresponding to the point cloud data and the three-dimensional bounding box corresponding to each target object.
  • the target object refers to the object whose bottom is the ground.
  • a target detection unit is stored in the computer device, and the 3D target detection unit can be used to directly perform target detection on the point cloud data, and identify the target object category corresponding to the point cloud data and the point cloud data corresponding to each target object.
  • the target object may be a vehicle, a pedestrian, or the like.
  • the bottom of the target object is the ground.
  • the computer device can calculate the three-dimensional bounding box of the corresponding target object according to the point cloud data corresponding to each target object, and the three-dimensional bounding box has corresponding center point coordinates, size, orientation, etc.
  • the method of target detection is any one of target detection methods such as shape matching-based 3D target detection, target local feature-based 3D target detection, and deep learning-based target detection methods.
  • Step 208 Extract the point cloud data in the three-dimensional bounding box, select target point cloud data from the extracted point cloud data, and perform plane fitting on the target point cloud data corresponding to the three-dimensional bounding box to obtain a fitted plane.
  • the target point cloud data refers to the plane fitting points used to fit the ground plane at the bottom of the target.
  • the computer equipment After obtaining the three-dimensional bounding box corresponding to each target object, the computer equipment extracts the number of point clouds in each three-dimensional bounding box, and selects the target point cloud data for the point cloud data extracted from each three-dimensional bounding box. Specifically, the computer device may perform grid division on the point cloud data extracted from each three-dimensional bounding box to obtain multiple grids. There can be multiple point cloud data corresponding to each grid. The computer equipment selects target point cloud data from the point cloud data corresponding to each grid, performs plane fitting on the target point cloud data corresponding to each three-dimensional bounding box, and obtains a fitting plane corresponding to each three-dimensional bounding box. The fitting plane is the ground plane at the bottom of the target object corresponding to the three-dimensional bounding box. The fitted plane is represented in the form of an equation.
  • the method of performing plane fitting on the target point cloud data corresponding to the three-dimensional bounding box may be any one of the least squares method, the eigenvalue method, the overall least squares method, and other plane fitting algorithms.
  • Step 210 extracting the position area corresponding to each target object according to the three-dimensional bounding box.
  • Step 212 Update the ground segmentation result according to the location area and the fitting plane to obtain the target ground segmentation result corresponding to the point cloud data.
  • the computer equipment can update the ground segmentation results through the fitting plane obtained by target detection. Yes, the computer device can determine the position area corresponding to each target object according to the three-dimensional bounding box obtained by the target detection.
  • the location area may be a horizontal plane area where the target object is located, that is, an area formed by the x-axis direction and the y-axis direction.
  • the computer equipment can further update the ground segmentation result according to the location area and the fitting plane, so as to update the less reliable data in the ground segmentation result, and obtain a more accurate target ground segmentation result.
  • updating the ground segmentation result according to the location area and the fitting plane, and obtaining the target ground segmentation result corresponding to the point cloud data includes: identifying the location relationship between each point in the point cloud data and the location area; When the point is not near the location area, keep the ground segmentation result corresponding to the point not near the location area in the ground segmentation result.
  • the computer equipment recognizes the positional relationship between each point in the point cloud data and the location area. Specifically, the computer equipment only needs to identify the positional relationship between the point and the location area according to the abscissa and ordinate of each point. When any one of the abscissa and ordinate is outside the location area, it means that the point is not near the location area. At this time, the computer device may retain the ground segmentation result corresponding to the point in the ground segmentation result, and use the category corresponding to the point in the ground segmentation result as the final category of the point. When both the abscissa and the ordinate of the point are within the location area, it indicates that the point is located near the location area. At this time, the computer device can calculate the distance between the point and the fitting plane, compare the distance with the threshold, update the ground segmentation result according to the comparison result, and then obtain the target ground segmentation result corresponding to the point cloud data.
  • the ground segmentation result is obtained by acquiring point cloud data and performing ground segmentation on the point cloud data. It can also perform target detection on the point cloud data to obtain the three-dimensional bounding box of the target object corresponding to the point cloud data. The target point cloud data corresponding to the bounding box is subjected to plane fitting to obtain the fitted plane. Then, the position area corresponding to each target object can be extracted according to the three-dimensional bounding box, and the ground segmentation result can be updated according to the position area and the fitting plane.
  • ground segmentation results are obtained by processing point cloud data through traditional ground segmentation methods, such as methods based on depth images and geometric relationships, normal vector methods, absolute height methods, and average height methods, it will be affected by noise, road slope, The influence of various factors such as weather is large, and it is only applicable to specific scenarios, which will reduce the accuracy of the ground segmentation results.
  • the target detection unit based on deep learning can adapt to a variety of scenarios, and is less affected by various factors such as noise, road slope, weather, etc., through the target detection unit based on deep learning stored in the computer device. Detect and obtain the fitting plane at the bottom of the target object, so as to update the ground segmentation result according to the fitting plane, which can improve the accuracy of the ground segmentation result.
  • the traditional ground segmentation method used for ground segmentation of point cloud data does not require manual annotation of ground data and training models, and uses a large number of sample data marked with ground data and deep neural network models for model training.
  • the trained model avoids the difficult and time-consuming problem of ground data labeling, and effectively improves the efficiency of point cloud data ground segmentation.
  • the step of selecting target point cloud data from the extracted point cloud data specifically includes:
  • Step 302 Divide the extracted point cloud data into multiple grids according to preset parameters.
  • Step 304 Select the point with the smallest height value in the point cloud data corresponding to each grid.
  • Step 306 Calculate the height difference between each point in the point cloud data corresponding to each grid and the point with the smallest height value.
  • Step 308 Select points whose height difference is less than the first threshold, and obtain target point cloud data according to the selected points.
  • the preset parameter may be a parameter for grid division of the data region where the extracted point cloud data is located.
  • the preset parameter may be length*width, indicating the length and width of each grid obtained after grid division. The length and width can be the same or different.
  • the preset parameters may also be equal division, the number of target grids, and the like.
  • the extracted point cloud data refers to the point cloud data in the three-dimensional bounding box corresponding to each target object.
  • the target point cloud data refers to the set of plane fitting points used to fit the ground plane at the bottom of the target.
  • the computer device may divide the data area corresponding to the extracted point cloud data in the x-axis direction and the y-axis direction according to preset parameters, thereby obtaining multiple grids.
  • the height of multiple grids is the same.
  • the order of division in the x-axis direction and the y-axis direction is not limited.
  • the computer device may first divide the data area corresponding to the extracted point cloud data in the x-axis direction according to preset parameters, and then divide the data area corresponding to the extracted point cloud data in the y-axis direction according to the preset parameters. It is also possible to first divide the data area corresponding to the extracted point cloud data in the y-axis direction according to the preset parameters, and then divide the data area corresponding to the extracted point cloud data in the x-axis direction according to the preset parameters.
  • the computer device selects the point with the smallest height value in the point cloud data corresponding to each grid, and calculates the height difference between each point in the point cloud data of the corresponding grid and the point with the smallest height value .
  • a first threshold for judging whether it is a plane fitting point is pre-stored in the computer device. When the height difference is less than the first threshold, it indicates that the point corresponding to the height difference is a plane fitting point.
  • the computer device compares the height difference with the first threshold, selects a point whose height difference is less than the first threshold, and determines the selected point as a plane fitting point, thereby obtaining target point cloud data.
  • the extracted point cloud data is divided into multiple grids according to preset parameters, and the point with the smallest height value is selected from the point cloud data corresponding to each grid, and the point cloud data corresponding to each grid is calculated The height difference between each point in the middle and the point with the smallest height value is selected, so that the point whose height difference is less than the first threshold is selected, and the target point cloud data is obtained according to the selected point. Since the grid division only needs to divide the data area corresponding to the point cloud data in the x-axis direction and the y-axis direction, and only needs to calculate the height between each point in the point cloud data corresponding to each grid and the point with the smallest height value The difference value can quickly select the target point cloud data. Therefore, in the unmanned mode, when the computing resources of the computer equipment are limited and the real-time requirements are high, the extraction efficiency of the target point cloud data can be improved.
  • performing plane fitting on the target point cloud data corresponding to the 3D bounding box to obtain the fitted plane includes: combining the target point cloud data of multiple grids corresponding to the 3D bounding box to obtain the 3D bounding box The corresponding combined point cloud data; perform plane fitting on the combined point cloud data of the three-dimensional bounding box to obtain the fitting plane corresponding to the three-dimensional bounding box.
  • the extracted point cloud data is divided into multiple grids, so each three-dimensional bounding box may correspond to multiple grids.
  • the target point cloud data is selected in each grid, the target point cloud data of multiple grids corresponding to each 3D bounding box can be combined together to obtain combined point cloud data corresponding to multiple 3D bounding boxes.
  • the computer equipment performs plane fitting on the combined point cloud data corresponding to each three-dimensional bounding box, to obtain a fitting plane corresponding to each three-dimensional bounding box.
  • the plane fitting method may be any one of the least squares method, the eigenvalue method, the overall least squares method and other plane fitting algorithms.
  • the fitting plane refers to the plane at the bottom of the target object corresponding to the 3D bounding box.
  • the fitting plane can be embodied in the form of a ternary linear equation.
  • the computer device combines the target point cloud data of multiple grids corresponding to each three-dimensional bounding box, and can combine the target point cloud data belonging to the same target object, which is beneficial to improve the calculation of each target.
  • the computer equipment performs plane fitting on the combined point cloud data corresponding to each three-dimensional bounding box to obtain the fitted plane corresponding to the three-dimensional bounding box, and can quickly calculate the bottom plane corresponding to each target object.
  • the steps of updating the ground segmentation result according to the location area and the fitting plane to obtain the target ground segmentation result specifically include:
  • Step 402 identifying the positional relationship between each point and the position area in the point cloud data.
  • Step 404 when the positional relationship is that the point is not near the location area, retain the ground segmentation result corresponding to the point not near the location area in the ground segmentation result.
  • Step 406 when the position relationship is that the point is located near the location area, calculate a first distance value between the point located near the location area and the fitting plane.
  • Step 408 Compare the first distance value with the second threshold, and determine the first category corresponding to the point located near the location area according to the comparison result.
  • Step 410 Update the ground segmentation result according to the first category to obtain the target ground segmentation result corresponding to the point cloud data.
  • the location area may be a horizontal plane area where the target object is located, that is, an area formed by the x-axis direction and the y-axis direction.
  • Computer equipment can identify the positional relationship between the point and the location area according to the abscissa and ordinate of each point in the point cloud data.
  • the computer device may retain the ground segmentation result corresponding to the point in the ground segmentation result, and use the category corresponding to the point in the ground segmentation result as the final category of the point.
  • the computer device can calculate the distance between the point and the fitting plane to obtain the first distance value.
  • a second threshold for judging the type of the point is pre-stored in the computer device. Therefore, the computer device compares the first distance with the second threshold, and determines the first category corresponding to the point according to the comparison result.
  • the first category refers to the category of points identified by the fitted plane.
  • the first category includes ground points and non-ground points
  • the first distance value is compared with the second threshold
  • determining the first category corresponding to the points located near the location area according to the comparison result includes: A point whose distance value is less than the second threshold is determined as a ground point; a point whose first distance value is greater than or equal to the second threshold is determined as a non-ground point.
  • the computer device determines that the first category corresponding to the point is a ground point.
  • the ground segmentation result is updated according to the first category of the point.
  • the computer device may retain the ground segmentation result corresponding to the point, and use the ground segmentation result corresponding to the point as the final ground segmentation result.
  • the ground segmentation result corresponding to the point not near the location area in the ground segmentation result is retained.
  • the distance between the point and the fitting plane is calculated, the distance is compared with the corresponding threshold, and the ground segmentation result is updated according to the comparison result. Since the fitting plane is obtained through detection processing by the target detection unit based on deep learning, the accuracy is high, and the ground segmentation results corresponding to the points near the location area can be updated according to the fitting plane, thereby improving the accuracy of the ground segmentation results. sex. At the same time, the problem of dividing points belonging to the surface of the same target object into multiple parts can also be avoided.
  • updating the ground segmentation results according to the first category includes: searching the ground segmentation results for the ground segmentation results corresponding to points located near the location area; comparing the found ground segmentation results with the first category Yes; when the comparisons are consistent, the found ground segmentation results are retained; when the comparisons are inconsistent, the found ground segmentation results are replaced according to the first category.
  • the first distance value between the point and the fitting plane can be calculated, and the first distance value corresponding to the point can be obtained by comparing the first distance value with the second threshold value.
  • a category Therefore, the computer equipment searches for the ground segmentation result corresponding to the point in the ground segmentation result, and identifies whether the found ground segmentation result is consistent with the first category corresponding to the calculated point. When the comparison is consistent, the computer equipment retains the ground of the point.
  • the segmentation result when the comparison is inconsistent, it is necessary to replace the ground segmentation result corresponding to the point with the first category, so as to update the ground segmentation result, and use the first category with higher accuracy as the ground segmentation result of the point. This improves the accuracy of the ground segmentation results.
  • performing ground segmentation on the point cloud data, and obtaining the ground segmentation result includes: dividing the point cloud area corresponding to the point cloud data into multiple sub-areas; calculating the point cloud data in each sub-area according to a preset plane equation The corresponding ground; calculate the second distance value between each point in the point cloud data of each sub-region and the corresponding ground; compare the second distance value with the third threshold, and determine the second category corresponding to each point according to the comparison result , and take the second category corresponding to each point as the ground segmentation result.
  • the point cloud area refers to the three-dimensional data space where the point cloud data is located.
  • the second distance value refers to the distance between each point in the point cloud data obtained by calculation and the corresponding ground in the process of dividing the point cloud data on the ground.
  • the third threshold refers to a threshold used to determine the category corresponding to each point in the point cloud data in the process of segmenting the point cloud data on the ground.
  • the second category refers to the category corresponding to each point in the ground segmentation result.
  • the computer equipment After acquiring the point cloud data, the computer equipment performs ground segmentation on the point cloud data. Specifically, the computer equipment divides the point cloud region corresponding to the point cloud data in the horizontal plane formed by the x-axis direction and the y-axis direction to obtain a plurality of sub-regions. Multiple points can be included in each subregion.
  • the division method can be equal division or random division.
  • the computer device may use the least squares method to estimate the corresponding ground according to the preset plane equation, so as to obtain the corresponding ground of each sub-region.
  • the ground corresponding to each sub-region is embodied in the form of a ternary linear equation.
  • the computer device traverses and inputs the point coordinates of each point in the corresponding sub-region in the equation corresponding to the ground, calculates the second distance between each point and the corresponding ground, and compares the second distance with the third threshold to obtain a comparison result.
  • the comparison result is that the second distance is smaller than the third threshold
  • the second category of the point corresponding to the comparison result is a ground point.
  • the comparison result is that the second distance is greater than or equal to the third threshold
  • the second category of the point corresponding to the comparison result is a non-ground point. Further, the second category corresponding to the multiple points is used as the ground segmentation result.
  • the computer equipment does not need to manually mark the ground data in the process of ground segmentation of the point cloud data, nor does it need to train the model according to the sample data marked with the ground data.
  • the data is segmented on the ground, which avoids the difficult and time-consuming problems of labeling, thereby improving the efficiency of the ground segmentation of point cloud data.
  • a ground segmentation device based on point cloud data including: a communication module 502, a ground segmentation module 504, a target detection module 506, a plane fitting module 508, and an extraction module 510 and update module 512, wherein:
  • the communication module 502 is used for acquiring point cloud data.
  • the ground segmentation module 504 is configured to perform ground segmentation on the point cloud data to obtain a ground segmentation result.
  • the target detection module 506 is configured to perform target detection on the point cloud data to obtain a three-dimensional bounding box of the target object corresponding to the point cloud data.
  • the plane fitting module 508 is used to extract the point cloud data in the three-dimensional bounding box, select the target point cloud data from the extracted point cloud data, and perform plane fitting on the target point cloud data corresponding to the three-dimensional bounding box to obtain a fitted plane .
  • the extraction module 510 is configured to extract the position area corresponding to each target object according to the three-dimensional bounding box.
  • the updating module 512 is configured to update the ground segmentation result according to the location area and the fitting plane to obtain the target ground segmentation result corresponding to the point cloud data.
  • the plane fitting module 508 is further configured to divide the extracted point cloud data into multiple grids according to preset parameters; select the point with the smallest height value from the point cloud data corresponding to each grid; calculate The height difference between each point in the point cloud data corresponding to each grid and the point with the smallest height value; select the point whose height difference is less than the first threshold, and obtain the target point cloud data according to the selected point.
  • the plane fitting module 508 is further configured to combine target point cloud data of multiple grids corresponding to the three-dimensional bounding box to obtain combined point cloud data corresponding to the three-dimensional bounding box;
  • the point cloud data is plane fitted to obtain the fitting plane corresponding to the 3D bounding box.
  • the update module 512 is further configured to identify the positional relationship between each point in the point cloud data and the location area; when the locational relationship is that the point is not near the location area, keep the ground segmentation results that are not near the location area The ground segmentation result corresponding to the point.
  • the above-mentioned device further includes:
  • the distance calculation module is configured to calculate the first distance value between the point located near the location area and the fitting plane when the location relationship is that the point is located near the location area.
  • the category determination module is configured to compare the first distance value with the second threshold, and determine the first category corresponding to the point located near the location area according to the comparison result.
  • the updating module 512 is further configured to update the ground segmentation result according to the first category to obtain the target ground segmentation result corresponding to the point cloud data.
  • the update module 512 is further configured to search the ground segmentation results for the ground segmentation results corresponding to the points located near the location area; compare the found ground segmentation results with the first category; when the comparison is consistent , keep the found ground segmentation results; when the comparison is inconsistent, replace the found ground segmentation results according to the first category.
  • the first category includes ground points and non-ground points
  • the category determination module is further configured to determine a point whose first distance value is less than a second threshold as a ground point; and determine that the first distance value is greater than or equal to the second Threshold points are identified as non-ground points.
  • the ground segmentation module 504 is further configured to divide the point cloud area corresponding to the point cloud data into a plurality of sub-areas; calculate the ground corresponding to the point cloud data of each sub-area according to a preset plane equation; calculate each sub-area The second distance value between each point in the point cloud data and the corresponding ground; compare the second distance value with the third threshold, determine the second category corresponding to each point according to the comparison result, category as the ground segmentation result.
  • Each module in the above-mentioned ground segmentation device based on point cloud data can be implemented in whole or in part by software, hardware and combinations thereof.
  • the above modules can be embedded in or independent of the processor in the computer device in the form of hardware, or stored in the memory in the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.
  • a computer device is provided, the internal structure of which can be shown in FIG. 6 .
  • the computer device includes a processor, memory, a communication interface, and a database connected by a system bus.
  • the processor of the computer device is used to provide computing and control capabilities.
  • the memory of the computer device includes a non-volatile storage medium, an internal memory.
  • the nonvolatile storage medium stores an operating system, a computer program, and a database.
  • the internal memory provides an environment for the execution of the operating system and computer programs in the non-volatile storage medium.
  • the database of the computer device is used to store point cloud data, ground segmentation results, target segmentation results, and the like.
  • the communication interface of the computer device is used for connection and communication with the vehicle sensor.
  • the computer program when executed by the processor, implements a ground segmentation method based on point cloud data.
  • FIG. 6 is only a block diagram of a partial structure related to the solution of the present application, and does not constitute a limitation on the computer equipment to which the solution of the present application is applied. Include more or fewer components than shown in the figures, or combine certain components, or have a different arrangement of components.
  • a computer device comprising a memory and one or more processors, the memory stores computer-readable instructions, and when the computer-readable instructions are executed by the one or more processors, makes the one or more processors execute the above methods to implement steps in the example.
  • One or more non-volatile computer-readable storage media storing computer-readable instructions, when the computer-readable instructions are executed by one or more processors, cause the one or more processors to execute the above method embodiments. step.
  • Nonvolatile memory may include read only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory.
  • Volatile memory may include random access memory (RAM) or external cache memory.
  • RAM is available in various forms such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous chain Road (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Length Measuring Devices By Optical Means (AREA)
  • Optical Radar Systems And Details Thereof (AREA)
  • Image Analysis (AREA)

Abstract

一种基于点云数据的地面分割方法，包括：获取点云数据；对点云数据进行地面分割，得到地面分割结果；对点云数据进行目标检测，得到点云数据对应的目标物体的三维包围框；提取三维包围框中的点云数据，在提取的点云数据中选取目标点云数据，将三维包围框对应的目标点云数据进行平面拟合，得到拟合平面；根据三维包围框提取各目标物体对应的位置区域；及根据位置区域以及拟合平面对地面分割结果进行更新，得到点云数据对应的目标地面分割结果。

Description

基于点云数据的地面分割方法、装置和计算机设备 技术领域
本申请涉及一种基于点云数据的地面分割方法、装置、计算机设备和存储介质。
背景技术
激光雷达传感器能够提供实时并且精确的三维场景信息，测距范围大、精度高，且测量值不受环境光照的影响，被应用于无人驾驶、安防监测、测绘勘探等领域。激光雷达传感器所采集的信息通常以点云的形式呈现，通过对点云数据进行相应地处理，可以用于完成高精度地图重建、障碍物的检测与跟踪等任务。由于地面点云数据在整个观测中通常占据了较大的数据量，且点云数据中包含的可用于完成目标任务的有用信息量较少，因此需要对点云数据进行地面分割，从而根据地面点云数据去除后的剩余点云数据来完成目标任务。
传统方式中，主要是通过地面拟合或者基于深度图像和几何关系的方法对点云数据进行地面分割。然而，这些方法只适用于特定场景下的点云数据地面分割，如地面拟合的方法只适用于道路较平坦的情况、基于深度图像和几何关系的方法需要保证激光雷达传感器的安装处于水平状态，当场景发生变化时，如果采用传统方式对点云数据进行地面分割，会导致点云数据的地面分割准确性较低。
发明内容
根据本申请公开的各种实施例，提供一种基于点云数据的地面分割方法、装置、计算机设备和存储介质。
一种基于点云数据的地面分割方法，包括：
获取点云数据；
对点云数据进行地面分割，得到地面分割结果；
对点云数据进行目标检测，得到点云数据对应的目标物体的三维包围框；
提取三维包围框中的点云数据，在提取的点云数据中选取目标点云数据，将三维包围框对应的目标点云数据进行平面拟合，得到拟合平面；
根据三维包围框提取各目标物体对应的位置区域；及
根据位置区域以及拟合平面对地面分割结果进行更新，得到点云数据对应的目标地面分割结果。
一种基于点云数据的地面分割装置，包括：
通信模块，用于获取点云数据；
地面分割模块，用于对点云数据进行地面分割，得到地面分割结果；
目标检测模块，用于对点云数据进行目标检测，得到点云数据对应的目标物体的三维包围框；
平面拟合模块，用于提取三维包围框中的点云数据，在提取的点云数据中选取目标点云数据，将三维包围框对应的目标点云数据进行平面拟合，得到拟合平面；
提取模块，用于根据三维包围框提取各目标物体对应的位置区域；及
更新模块，用于根据位置区域以及拟合平面对地面分割结果进行更新，得到点云数据对应的目标地面分割结果。
一种计算机设备，包括存储器和一个或多个处理器，存储器中储存有计算机可读指令，计算机可读指令被处理器执行时，使得一个或多个处理器执行以下步骤：
获取点云数据；
对点云数据进行地面分割，得到地面分割结果；
对点云数据进行目标检测，得到点云数据对应的目标物体的三维包围框；
提取三维包围框中的点云数据，在提取的点云数据中选取目标点云数据，将三维包围框对应的目标点云数据进行平面拟合，得到拟合平面；
根据三维包围框提取各目标物体对应的位置区域；及
根据位置区域以及拟合平面对地面分割结果进行更新，得到点云数据对应的目标地面分割结果。
一个或多个存储有计算机可读指令的非易失性计算机可读存储介质，计算机可读指令被一个或多个处理器执行时，使得一个或多个处理器执行上述以下步骤：
获取点云数据；
对点云数据进行地面分割，得到地面分割结果；
对点云数据进行目标检测，得到点云数据对应的目标物体的三维包围框；
提取三维包围框中的点云数据，在提取的点云数据中选取目标点云数据，将三维包围框对应的目标点云数据进行平面拟合，得到拟合平面；
根据三维包围框提取各目标物体对应的位置区域；及
根据位置区域以及拟合平面对地面分割结果进行更新，得到点云数据对应的目标地面分割结果。
本申请的一个或多个实施例的细节在下面的附图和描述中提出。本申请的其它特征和优点将从说明书、附图以及权利要求书变得明显。
附图说明
为了更清楚地说明本申请实施例中的技术方案，下面将对实施例中所需要使用的附图作简单地介绍，显而易见地，下面描述中的附图仅仅是本申请的一些实施例，对于本领域普通技术人员来讲，在不付出创造性劳动的前提下，还可以根据这些附图获得其它的附图。
图1为一个或多个实施例中基于点云数据的地面分割方法的应用场景图。
图2为一个或多个实施例中基于点云数据的地面分割方法的流程示意图。
图3为一个或多个实施例中在提取的点云数据中选取目标点云数据步骤的流程示意图。
图4为一个或多个实施例中根据位置区域以及拟合平面对地面分割结果进行更新，得到 目标地面分割结果步骤的流程示意图。
图5为一个或多个实施例中基于点云数据的地面分割装置的框图。
图6为一个或多个实施例中计算机设备的框图。
具体实施方式
为了使本申请的技术方案及优点更加清楚明白，以下结合附图及实施例，对本申请进行进一步详细说明。应当理解，此处描述的具体实施例仅仅用以解释本申请，并不用于限定本申请。
需要说明的是，本申请的说明书和权利要求书中的术语“第一”、“第二”等是用于区别类似的对象，而不必用于描述特定的顺序或先后次序。
本申请提供的基于点云数据的地面分割方法，可以应用于如图1所示的无人驾驶的应用场景中。无人驾驶车辆中预先安装有车载传感器102以及车载计算机设备104。车载计算机设备可以简称为计算机设备。车载传感器102与计算机设备104进行通信，在无人驾驶过程中，车载传感器102将采集到的点云数据传送至计算机设备104。计算机设备104对点云数据进行地面分割，得到地面分割结果。计算机设备104还可以对点云数据进行目标检测，得到点云数据对应的目标物体的三维包围框。计算机设备104提取三维包围框中的点云数据，在提取的点云数据中选取目标点云数据，将三维包围框对应的目标点云数据进行平面拟合，得到拟合平面。计算机设备104根据三维包围框提取各目标物体对应的位置区域，从而根据位置区域以及拟合平面对地面分割结果进行更新，得到点云数据对应的目标地面分割结果。车载传感器102可以是激光雷达。
在其中一个实施例中，如图2所示，提供了一种基于点云数据的地面分割方法，以该方法应用于图1中的计算机设备为例进行说明，包括以下步骤：
步骤202，获取点云数据。
点云数据可以是车载传感器将扫描到的周围环境信息以点云形式记录的数据。点云数据具体可以包括各点的三维坐标(x，y，z)、激光反射强度(Intensity)、颜色信息(RGB)等。三维坐标用于表示周围环境中目标物体表面的位置信息。例如，三维坐标可以是点在笛卡尔坐标系中的坐标，具体包括点在笛卡尔坐标系中的横轴坐标、纵轴坐标和竖轴坐标。笛卡尔坐标系是以车载传感器为原点建立的三维空间坐标系，三维空间坐标系包括横轴(x轴)、纵轴(y轴)和竖轴(z轴)。以车载传感器为原点建立的三维空间坐标系满足右手定则。三维坐标中的x轴坐标表示扫描到的目标物体表面相对于车载传感器的纵向距离，y轴坐标表示扫描到的目标物体表面相对于车载传感器的横向偏移量，z轴坐标表示扫描到的目标物体表面相对于车载传感器的高度。
车辆在无人驾驶过程中，可以通过安装在车辆上的车载传感器对当前环境进行扫描，得到相应的点云数据，并将采集到的点云数据传送至计算机设备。例如，车载传感器可以是激光雷达。
步骤204，对点云数据进行地面分割，得到地面分割结果。
地面分割是指识别点云数据中的地面点与非地面点。由于点云数据中包括大量的地面点云数据，且点云数据中包含的可用于完成高精度地图重建、障碍物的检测与跟踪等任务的有用信息量较少，因此需要对点云数据进行地面分割，从而根据地面点去除后的剩余点云数据来完成目标任务。
计算机设备可以对点云数据进行地面分割，从而识别点云数据中的各点的类别，得到相应的地面分割结果。类别具体包括地面点和非地面点。地面分割的方式可以是先将点云数据所在的点云区域划分为多个子区域。点云区域是指点云数据所在的三维数据空间。划分方式可以是对点云区域进行栅格划分，即对点云区域在x轴方向与y轴方向形成的水平面进行划分，栅格划分方式可以是均等划分，也可以是随机划分。例如，对于可视范围为100m的车载传感器，车载传感器扫描区域内的水平面大小为100m*100m，则可将点云区域均等划分成10*10个水平格子。针对划分得到的每个子区域，计算机设备可以根据预设平面方程采用最小二乘法估计相应的地面，从而得到每个子区域对应的地面。例如，预设平面方程可以是三元一次方程。每个子区域对应的地面是以三元一次方程的形式体现的。计算机设备在地面对应的方程中遍历输入相应子区域中的点坐标，计算各点与相应的地面之间的距离，当距离小于阈值时，则将该点确定为地面点。当距离大于或者等于阈值时，则将该点确定为非地面点。阈值是指用于判断该点是否为地面点的距离阈值。
在其中一个实施例中，计算机设备还可以采用基于深度图像和几何关系的方法、法向量方法、绝对高度方法、平均高度方法等不需要标准地面数据和训练模型的点云地面分割方法中的任意一种对点云数据进行地面分割。
步骤206，对点云数据进行目标检测，得到点云数据对应的目标物体的三维包围框。
目标检测是指识别点云数据对应的目标物体类别以及各目标物体对应的三维包围框。目标物体是指目标底部为地面的物体。
计算机设备中存储有目标检测单元，可以直接利用三维目标检测单元对点云数据进行目标检测，识别点云数据对应的目标物体类别以及各目标物体对应的点云数据。例如，目标物体可以是车辆、行人等。该目标物体的底部为地面。计算机设备可以根据各目标物体对应的点云数据计算相应目标物体的三维包围框，三维包围框存在对应的中心点坐标、大小、朝向等。目标检测的方式是基于形状匹配的三维目标检测、基于目标局部特征的三维目标检测、基于深度学习的目标检测方法等目标检测方法中的任意一种。
步骤208，提取三维包围框中的点云数据，在提取的点云数据中选取目标点云数据，将三维包围框对应的目标点云数据进行平面拟合，得到拟合平面。
目标点云数据是指用于拟合目标底部的地平面的平面拟合点。
计算机设备在得到各目标物体对应的三维包围框后，将每个三维包围框中的点云数提取出来，针对每个三维包围框提取出来的点云数据进行目标点云数据的选取。具体的，计算机设备可以对每个三维包围框中提取的点云数据进行栅格划分，得到多个栅格。每个栅格中可以对应存在多个点云数据。计算机设备在每个栅格对应的点云数据中选取目标点云数据，将每个三维包围框对应的目标点云数据进行平面拟合，得到每个三维包围框对应的拟合平面。 拟合平面为三维包围框对应的目标物体底部的地平面。拟合平面是以方程的形式体现的。
在一个实施例中，将三维包围框对应的目标点云数据进行平面拟合的方式可以是最小二乘法、特征值法、总体最小二乘法等平面拟合算法中的任意一种。
步骤210，根据三维包围框提取各目标物体对应的位置区域。
步骤212，根据位置区域以及拟合平面对地面分割结果进行更新，得到点云数据对应的目标地面分割结果。
由于地面分割结果会受到噪音、道路坡度、天气等多种因素的影响，会导致地面分割结果的准确性降低，因此，计算机设备可以通过目标检测得到的拟合平面对地面分割结果进行更新，具体的，计算机设备可以根据目标检测得到的三维包围框确定每个目标物体对应的位置区域。位置区域可以是目标物体所在的水平面区域，即x轴方向与y轴方向所组成的区域。计算机设备进而可以根据位置区域以及拟合平面对地面分割结果进行更新，从而将地面分割结果中可靠性较低的数据进行更新，得到更为准确的目标地面分割结果。
在其中一个实施例中，根据位置区域以及拟合平面对地面分割结果进行更新，得到点云数据对应的目标地面分割结果包括：识别点云数据中各点与位置区域的位置关系；当位置关系为点未在位置区域附近时，保留地面分割结果中未在位置区域附近的点对应的地面分割结果。
计算机设备识别点云数据中的每个点与位置区域的位置关系，具体的，计算机设备只需要根据每个点的横坐标以及纵坐标来识别该点与位置区域的位置关系，当该点的横坐标以及纵坐标中任意一个坐标位于位置区域外时，则表明该点未在位置区域附近。此时，计算机设备可以保留地面分割结果中该点对应的地面分割结果，将地面分割结果中该点对应的类别作为点的最终类别。当该点的横坐标与纵坐标均在位置区域内时，则表明该点位于位置区域附近。此时，计算机设备可以计算该点与拟合平面之间的距离，将距离与阈值进行比较，根据比较结果对地面分割结果进行更新，进而得到点云数据对应的目标地面分割结果。
在本实施例中，通过获取点云数据，对点云数据进行地面分割，得到地面分割结果。还可以对点云数据进行目标检测，得到点云数据对应的目标物体的三维包围框，通过提取三维包围框中的点云数据，在提取的点云数据中选取目标点云数据，从而将三维包围框对应的目标点云数据进行平面拟合，得到拟合平面。进而可以根据三维包围框提取各目标物体对应的位置区域，并根据位置区域以及拟合平面对地面分割结果进行更新。由于地面分割结果是通过传统的地面分割方式，如基于深度图像和几何关系的方法、法向量方法、绝对高度方法、平均高度方法等对点云数据进行处理得到的，会受到噪音、道路坡度、天气等多种因素的较大影响，且只适用于特定场景，会导致地面分割结果的准确性降低。而基于深度学习的目标检测单元能够适应多种场景，且受噪音、道路坡度、天气等多种因素的影响较小，通过计算机设备中存储的基于深度学习的目标检测单元对点云数据进行目标检测，获取目标物体底部的拟合平面，从而根据拟合平面对地面分割结果进行更新，能够提高地面分割结果的准确性。另外，对点云数据进行地面分割时所采用的传统地面分割方式是不需要进行人工标注地面数据以及训练模型的，与利用大量标注有地面数据的样本数据以及深度神经网络模型进行模型 训练，通过训练后的模型进行点云数据地面分割的方式相比，避免了地面数据标注难度较大且耗时的问题，有效提高了点云数据地面分割的效率。
在其中一个实施例中，如图3所示，在提取的点云数据中选取目标点云数据的步骤，具体包括：
步骤302，根据预设参数将提取的点云数据划分为多个栅格。
步骤304，在各栅格对应的点云数据中选取高度值最小的点。
步骤306，计算各栅格对应的点云数据中各点与高度值最小的点之间的高度差值。
步骤308，选取高度差值小于第一阈值的点，根据选取的点得到目标点云数据。
预设参数可以是对提取的点云数据所在的数据区域进行栅格划分的参数。例如，预设参数可以是长*宽，表明栅格划分后得到的每个栅格的长度和宽度。长和宽可以是相同的，也可以是不同的。预设参数还可以是均等划分、目标栅格数量等。提取的点云数据是指每个目标物体对应的三维包围框中的点云数据。目标点云数据是指用于拟合目标底部的地平面的平面拟合点的集合。
具体的，计算机设备可以根据预设参数将提取的点云数据对应的数据区域进行x轴方向以及y轴方向的划分，从而得到多个栅格。多个栅格的高度是相同的。对提取的点云数据对应的数据区域进行栅格划分时，x轴方向以及y轴方向划分的顺序不作限定。例如，计算机设备可以先根据预设参数对提取的点云数据对应的数据区域进行x轴方向的划分，再根据预设参数将提取的点云数据对应的数据区域进行y轴方向的划分。也可以先根据预设参数对提取的点云数据对应的数据区域进行y轴方向的划分，再根据预设参数将提取的点云数据对应的数据区域进行x轴方向的划分。
针对每个栅格，计算机设备在每个栅格对应的点云数据中选取高度值最小的点，并计算相应栅格的点云数据中各点与高度值最小的点之间的高度差值。计算机设备中预先存储有用于判断是否为平面拟合点的第一阈值。当高度差值小于第一阈值时，表明该高度差值对应的点为平面拟合点。计算机设备将高度差值与第一阈值进行比较，选取高度差值小于第一阈值的点，将选取的点确定为平面拟合点，从而得到目标点云数据。
在本实施例中，根据预设参数将提取的点云数据划分为多个栅格，并在各栅格对应的点云数据中选取高度值最小的点，计算各栅格对应的点云数据中各点与高度值最小的点之间的高度差值，从而选取高度差值小于第一阈值的点，根据选取的点得到目标点云数据。由于栅格划分只需要对点云数据对应的数据区域进行x轴方向以及y轴方向的划分，且只需要计算各栅格对应的点云数据中各点与高度值最小的点之间的高度差值，即可快速选取目标点云数据，因此，能够在无人驾驶模式下，计算机设备的计算资源有限且实时性要求高时，提高目标点云数据的提取效率。
在其中一个实施例中，将三维包围框对应的目标点云数据进行平面拟合，得到拟合平面包括：将三维包围框对应的多个栅格的目标点云数据进行组合，得到三维包围框对应的组合点云数据；将三维包围框的组合点云数据进行平面拟合，得到三维包围框对应的拟合平面。
由于将三维包围框中的点云数据提取出来之后，对提取出来的点云数据划分为多个栅格， 因此每个三维包围框可以对应多个栅格。在每个栅格中选取目标点云数据后，可以将每个三维包围框对应的多个栅格的目标点云数据组合在一起，得到多个三维包围框对应的组合点云数据。进而计算机设备对每个三维包围框对应的组合点云数据分别进行平面拟合，得到每个三维包围框对应的拟合平面。平面拟合的方式可以是最小二乘法、特征值法、总体最小二乘法等平面拟合算法中的任意一种。拟合平面是指三维包围框对应的目标物体底部的平面。拟合平面可以是以三元一次方程的形式体现的。
在本实施例中，计算机设备将每个三维包围框对应的多个栅格的目标点云数据进行组合，能够将属于同一目标物体的目标点云数据组合在一起，有利于提高计算每个目标物体底部的拟合平面的效率。计算机设备对每个三维包围框对应的组合点云数据进行平面拟合，得到三维包围框对应的拟合平面，能够快速计算得到每个目标物体所对应的底部平面。
在其中一个实施例中，如图4所示，根据位置区域以及拟合平面对地面分割结果进行更新，得到目标地面分割结果的步骤，具体包括：
步骤402，识别点云数据中各点与位置区域的位置关系。
步骤404，当位置关系为点未在位置区域附近时，保留地面分割结果中未在位置区域附近的点对应的地面分割结果。
步骤406，当位置关系为点位于位置区域附近时，计算位于位置区域附近的点与拟合平面之间的第一距离值。
步骤408，将第一距离值与第二阈值进行比较，根据比较结果确定位于位置区域附近的点对应的第一类别。
步骤410，根据第一类别对地面分割结果进行更新，得到点云数据对应的目标地面分割结果。
位置区域可以是目标物体所在的水平面区域，即x轴方向与y轴方向所组成的区域。
计算机设备可以根据点云数据中各点的横坐标以及纵坐标来识别该点与位置区域的位置关系，当点的横坐标以及纵坐标中任意一个坐标位于位置区域外时，则表明该点未在位置区域附近，该点不属于相应的三维包围框对应的目标物体表面的点。此时，计算机设备可以保留地面分割结果中该点对应的地面分割结果，将地面分割结果中该点对应的类别作为点的最终类别。当点的横坐标与纵坐标均在位置区域内时，则表明该点位于位置区域附近，该点属于相应的三维包围框对应的目标物体表面的点。由于位置区域附近的点的分割结果准确性容易受多种因素的影响。此时，计算机设备可以计算该点与拟合平面之间的距离，得到第一距离值。计算机设备中预先存储有用于判断点的类别的第二阈值。因此，计算机设备将第一距离与第二阈值进行比较，根据比较结果确定点对应的第一类别。第一类别是指通过拟合平面识别得到的点的类别。
在其中一个实施例中，第一类别包括地面点以及非地面点，将第一距离值与第二阈值进行比较，根据比较结果确定位于位置区域附近的点对应的第一类别包括：将第一距离值小于第二阈值的点确定为地面点；将第一距离值大于或者等于第二阈值的点确定为非地面点。
当比较结果为第一距离小于第二阈值时，计算机设备确定点对应的第一类别为地面点。从而根据该点的第一类别对地面分割结果进行更新。当比较结果为第一距离大于或者等于第二阈值时，计算机设备可以保留该点对应的地面分割结果，并将该点对应的地面分割结果作为最终的地面分割结果。
在本实施例中，通过识别点云数据中各点是否位于位置区域附近，当点未在位置区域附件时，则保留地面分割结果中未在位置区域附近的点对应的地面分割结果。当点位于位置区域附近时，计算该点与拟合平面之间的距离，将距离与相应的阈值进行比较，根据比较结果对地面分割结果进行更新。由于拟合平面是基于深度学习的目标检测单元进行检测处理得到的，准确性较高，由于可以根据拟合平面来更新位置区域附近的点对应的地面分割结果，从而提高了地面分割结果的准确性。同时，还能够避免将属于同一目标物体表面的点划分为多个部分的问题。
在其中一个实施例中，根据第一类别对地面分割结果进行更新包括：在地面分割结果中查找位于位置区域附近的点对应的地面分割结果；将查找到的地面分割结果与第一类别进行比对；当比对一致时，保留查找到的地面分割结果；当比对不一致时，根据第一类别对查找到的地面分割结果进行更换。
当点云数据中存在位于位置区域附近的点时，可以计算该点与拟合平面之间的第一距离值，并通过将第一距离值与第二阈值进行比较，得到该点对应的第一类别。从而计算机设备在地面分割结果中查找该点对应的地面分割结果，识别查找到的地面分割结果与计算得到的点对应的第一类别是否一致，当比对一致时，计算机设备保留该点的地面分割结果，当比对不一致时，则需要将该点对应的地面分割结果更换为第一类别，从而对地面分割结果进行更新，采用准确度更高的第一类别作为该点的地面分割结果，进而提高了地面分割结果的准确性。
在其中一个实施例中，对点云数据进行地面分割，得到地面分割结果包括：将点云数据对应的点云区域划分为多个子区域；根据预设平面方程计算各子区域中的点云数据对应的地面；计算各子区域的点云数据中各点与相应的地面之间的第二距离值；将第二距离值与第三阈值进行比较，根据比较结果确定各点对应的第二类别，将各点对应的第二类别作为地面分割结果。
点云区域是指点云数据所在的三维数据空间。第二距离值是指对点云数据进行地面分割的过程中，计算得到的点云数据中各点与相应的地面之间的距离。第三阈值是指对点云数据进行地面分割的过程中，用于判断点云数据中各点对应的类别的阈值。第二类别是指地面分割结果中各点对应的类别。
计算机设备在获取到点云数据后，对点云数据进行地面分割。具体的，计算机设备对点云数据对应的点云区域在x轴方向与y轴方向形成的水平面进行划分，得到多个子区域。每个子区域中可以包括多个点。划分方式可以是均等划分，也可以是随机划分。针对划分得到的每个子区域，计算机设备可以根据预设平面方程采用最小二乘法估计相应的地面，从而得到每个子区域对应的地面。每个子区域对应的地面是以三元一次方程的形式体现的。计算机 设备在地面对应的方程中遍历输入相应子区域中各点的点坐标，计算各点与相应的地面之间的第二距离，并将第二距离与第三阈值进行比较，得到比较结果。当比较结果为第二距离小于第三阈值时，则该比较结果对应的点的第二类别为地面点。当比较结果为第二距离大于或者等于第三阈值时，则该比较结果对应的点的第二类别为非地面点。进而将多个点对应的第二类别作为地面分割结果。
在本实施例中，计算机设备在对点云数据进行地面分割的过程中，并不需要人工标注地面数据，也不需要根据标注有地面数据的样本数据训练模型，根据训练后的模型对点云数据进行地面分割，避免了标注难度较大，且耗时费力的问题，由此提高了点云数据地面分割的效率。
在其中一个实施例中，如图5所示，提供了一种基于点云数据的地面分割装置，包括：通信模块502、地面分割模块504、目标检测模块506、平面拟合模块508、提取模块510和更新模块512，其中：
通信模块502，用于获取点云数据。
地面分割模块504，用于对点云数据进行地面分割，得到地面分割结果。
目标检测模块506，用于对点云数据进行目标检测，得到点云数据对应的目标物体的三维包围框。
平面拟合模块508，用于提取三维包围框中的点云数据，在提取的点云数据中选取目标点云数据，将三维包围框对应的目标点云数据进行平面拟合，得到拟合平面。
提取模块510，用于根据三维包围框提取各目标物体对应的位置区域。
更新模块512，用于根据位置区域以及拟合平面对地面分割结果进行更新，得到点云数据对应的目标地面分割结果。
在其中一个实施例中，平面拟合模块508还用于根据预设参数将提取的点云数据划分为多个栅格；在各栅格对应的点云数据中选取高度值最小的点；计算各栅格对应的点云数据中各点与高度值最小的点之间的高度差值；选取高度差值小于第一阈值的点，根据选取的点得到目标点云数据。
在其中一个实施例中，平面拟合模块508还用于将三维包围框对应的多个栅格的目标点云数据进行组合，得到三维包围框对应的组合点云数据；将三维包围框的组合点云数据进行平面拟合，得到三维包围框对应的拟合平面。
在其中一个实施例中，更新模块512还用于识别点云数据中各点与位置区域的位置关系；当位置关系为点未在位置区域附近时，保留地面分割结果中未在位置区域附近的点对应的地面分割结果。
在其中一个实施例中，上述装置还包括：
距离计算模块，用于当位置关系为点位于位置区域附近时，计算位于位置区域附近的点与拟合平面之间的第一距离值。
类别确定模块，用于将第一距离值与第二阈值进行比较，根据比较结果确定位于位置区域附近的点对应的第一类别。
更新模块512还用于根据第一类别对地面分割结果进行更新，得到点云数据对应的目标地面分割结果。
在其中一个实施例中，更新模块512还用于在地面分割结果中查找位于位置区域附近的点对应的地面分割结果；将查找到的地面分割结果与第一类别进行比对；当比对一致时，保留查找到的地面分割结果；当比对不一致时，根据第一类别对查找到的地面分割结果进行更换。
在其中一个实施例中，第一类别包括地面点以及非地面点，类别确定模块还用于将第一距离值小于第二阈值的点确定为地面点；将第一距离值大于或者等于第二阈值的点确定为非地面点。
在其中一个实施例中，地面分割模块504还用于将点云数据对应的点云区域划分为多个子区域；根据预设平面方程计算各子区域的点云数据对应的地面；计算各子区域的点云数据中各点与相应的地面之间的第二距离值；将第二距离值与第三阈值进行比较，根据比较结果确定各点对应的第二类别，将各点对应的第二类别作为地面分割结果。
关于基于点云数据的地面分割装置的具体限定可以参见上文中对于基于点云数据的地面分割方法的限定，在此不再赘述。上述基于点云数据的地面分割装置中的各个模块可全部或部分通过软件、硬件及其组合来实现。上述各模块可以硬件形式内嵌于或独立于计算机设备中的处理器中，也可以以软件形式存储于计算机设备中的存储器中，以便于处理器调用执行以上各个模块对应的操作。
在一个实施例中，提供了一种计算机设备，其内部结构图可以如图6所示。该计算机设备包括通过系统总线连接的处理器、存储器、通信接口和数据库。其中，该计算机设备的处理器用于提供计算和控制能力。该计算机设备的存储器包括非易失性存储介质、内存储器。该非易失性存储介质存储有操作系统、计算机程序和数据库。该内存储器为非易失性存储介质中的操作系统和计算机程序的运行提供环境。该计算机设备的数据库用于存储点云数据、地面分割结果、目标分割结果等。该计算机设备的通信接口用于与车载传感器进行连接通信。该计算机程序被处理器执行时以实现一种基于点云数据的地面分割方法。
本领域技术人员可以理解，图6中示出的结构，仅仅是与本申请方案相关的部分结构的框图，并不构成对本申请方案所应用于其上的计算机设备的限定，具体的计算机设备可以包括比图中所示更多或更少的部件，或者组合某些部件，或者具有不同的部件布置。
一种计算机设备，包括存储器及一个或多个处理器，存储器中储存有计算机可读指令，计算机可读指令被一个或多个处理器执行时，使得一个或多个处理器执行上述各个方法实施例中的步骤。
一个或多个存储有计算机可读指令的非易失性计算机可读存储介质，计算机可读指令被一个或多个处理器执行时，使得一个或多个处理器执行上述各个方法实施例中的步骤。
本领域普通技术人员可以理解实现上述实施例方法中的全部或部分流程，是可以通过计算机可读指令来指令相关的硬件来完成，所述的计算机可读指令可存储于一非易失性计算机可读取存储介质中，该计算机可读指令在执行时，可包括如上述各方法的实施例的流程。其 中，本申请所提供的各实施例中所使用的对存储器、存储、数据库或其它介质的任何引用，均可包括非易失性和/或易失性存储器。非易失性存储器可包括只读存储器(ROM)、可编程ROM(PROM)、电可编程ROM(EPROM)、电可擦除可编程ROM(EEPROM)或闪存。易失性存储器可包括随机存取存储器(RAM)或者外部高速缓冲存储器。作为说明而非局限，RAM以多种形式可得，诸如静态RAM(SRAM)、动态RAM(DRAM)、同步DRAM(SDRAM)、双数据率SDRAM(DDRSDRAM)、增强型SDRAM(ESDRAM)、同步链路(Synchlink)DRAM(SLDRAM)、存储器总线(Rambus)直接RAM(RDRAM)、直接存储器总线动态RAM(DRDRAM)、以及存储器总线动态RAM(RDRAM)等。
以上实施例的各技术特征可以进行任意的组合，为使描述简洁，未对上述实施例中的各个技术特征所有可能的组合都进行描述，然而，只要这些技术特征的组合不存在矛盾，都应当认为是本说明书记载的范围。
以上所述实施例仅表达了本申请的几种实施方式，其描述较为具体和详细，但并不能因此而理解为对发明专利范围的限制。应当指出的是，对于本领域的普通技术人员来说，在不脱离本申请构思的前提下，还可以做出若干变形和改进，这些都属于本申请的保护范围。因此，本申请专利的保护范围应以所附权利要求为准。

Claims (20)

  1. 一种基于点云数据的地面分割方法，包括：
    获取点云数据；
    对所述点云数据进行地面分割，得到地面分割结果；
    对所述点云数据进行目标检测，得到所述点云数据对应的目标物体的三维包围框；
    提取所述三维包围框中的点云数据，在提取的点云数据中选取目标点云数据，将所述三维包围框对应的目标点云数据进行平面拟合，得到拟合平面；
    根据所述三维包围框提取各目标物体对应的位置区域；及
    根据所述位置区域以及所述拟合平面对所述地面分割结果进行更新，得到所述点云数据对应的目标地面分割结果。
  2. 根据权利要求1所述的方法，其特征在于，所述在提取的点云数据中选取目标点云数据包括：
    根据预设参数将提取的点云数据划分为多个栅格；
    在各栅格对应的点云数据中选取高度值最小的点；
    计算各栅格对应的点云数据中各点与所述高度值最小的点之间的高度差值；及
    选取高度差值小于第一阈值的点，根据选取的点得到目标点云数据。
  3. 根据权利要求2所述的方法，其特征在于，所述将所述三维包围框对应的目标点云数据进行平面拟合，得到拟合平面包括：
    将所述三维包围框对应的多个栅格的目标点云数据进行组合，得到所述三维包围框对应的组合点云数据；及
    将所述三维包围框的组合点云数据进行平面拟合，得到所述三维包围框对应的拟合平面。
  4. 根据权利要求1所述的方法，其特征在于，所述根据所述位置区域以及所述拟合平面对所述地面分割结果进行更新，得到目标地面分割结果包括：
    识别所述点云数据中各点与所述位置区域的位置关系；及
    当所述位置关系为点未在所述位置区域附近时，保留所述地面分割结果中未在所述位置区域附近的点对应的地面分割结果。
  5. 根据权利要求4所述的方法，其特征在于，所述方法还包括：
    当所述位置关系为点位于所述位置区域附近时，计算位于所述位置区域附近的点与所述拟合平面之间的第一距离值；
    将所述第一距离值与第二阈值进行比较，根据比较结果确定所述位于所述位置区域附近的点对应的第一类别；及
    根据所述第一类别对所述地面分割结果进行更新，得到所述点云数据对应的目标地面分割结果。
  6. 根据权利要求5所述的方法，其特征在于，所述根据所述第一类别对所述地面分割结果进行更新包括：
    在所述地面分割结果中查找所述位于所述位置区域附近的点对应的地面分割结果；
    将查找到的地面分割结果与所述第一类别进行比对；
    当比对一致时，保留所述查找到的地面分割结果；及
    当比对不一致时，根据所述第一类别对所述查找到的地面分割结果进行更换。
  7. 根据权利要求5所述的方法，其特征在于，所述第一类别包括地面点以及非地面点，所述将所述第一距离值与第二阈值进行比较，根据比较结果确定所述位于所述位置区域附近的点对应的第一类别包括：
    将所述第一距离值小于第二阈值的点确定为地面点；及
    将所述第一距离值大于或者等于第二阈值的点确定为非地面点。
  8. 根据权利要求1至7任意一项所述的方法，其特征在于，所述对所述点云数据进行地面分割，得到地面分割结果包括：
    将所述点云数据对应的点云区域划分为多个子区域；
    根据预设平面方程计算各子区域的点云数据对应的地面；
    计算各子区域的点云数据中各点与相应的地面之间的第二距离值；及
    将第二距离值与第三阈值进行比较，根据比较结果确定各点对应的第二类别，将各点对应的第二类别作为地面分割结果。
  9. 一种基于点云数据的地面分割装置，包括：
    通信模块，用于获取点云数据；
    地面分割模块，用于对所述点云数据进行地面分割，得到地面分割结果；
    目标检测模块，用于对所述点云数据进行目标检测，得到所述点云数据对应的目标物体的三维包围框；
    平面拟合模块，用于提取所述三维包围框中的点云数据，在提取的点云数据中选取目标点云数据，将所述三维包围框对应的目标点云数据进行平面拟合，得到拟合平面；
    提取模块，用于根据所述三维包围框提取各目标物体对应的位置区域；及
    更新模块，用于根据所述位置区域以及所述拟合平面对所述地面分割结果进行更新，得到所述点云数据对应的目标地面分割结果。
  10. 根据权利要求9所述的装置，其特征在于，所述平面拟合模块还用于根据预设参数将提取的点云数据划分为多个栅格；在各栅格对应的点云数据中选取高度值最小的点；计算各栅格对应的点云数据中各点与所述高度值最小的点之间的高度差值；及选取高度差值小于第一阈值的点，根据选取的点得到目标点云数据。
  11. 一种计算机设备，包括存储器及一个或多个处理器，所述存储器中存储有计算机可读指令，所述计算机可读指令被所述一个或多个处理器执行时，使得所述一个或多个处理器执行以下步骤：
    获取点云数据；
    对所述点云数据进行地面分割，得到地面分割结果；
    对所述点云数据进行目标检测，得到所述点云数据对应的目标物体的三维包围框；
    提取所述三维包围框中的点云数据，在提取的点云数据中选取目标点云数据，将所述三维包围框对应的目标点云数据进行平面拟合，得到拟合平面；
    根据所述三维包围框提取各目标物体对应的位置区域；及
    根据所述位置区域以及所述拟合平面对所述地面分割结果进行更新，得到所述点云数据对应的目标地面分割结果。
  12. 根据权利要求11所述的计算机设备，其特征在于，所述处理器执行所述计算机可读指令时还执行以下步骤：根据预设参数将提取的点云数据划分为多个栅格；在各栅格对应的点云数据中选取高度值最小的点；计算各栅格对应的点云数据中各点与所述高度值最小的点之间的高度差值；及选取高度差值小于第一阈值的点，根据选取的点得到目标点云数据。
  13. 根据权利要求12所述的计算机设备，其特征在于，所述处理器执行所述计算机可读指令时还执行以下步骤：将所述三维包围框对应的多个栅格的目标点云数据进行组合，得到所述三维包围框对应的组合点云数据；及将所述三维包围框的组合点云数据进行平面拟合，得到所述三维包围框对应的拟合平面。
  14. 根据权利要求11所述的计算机设备，其特征在于，所述处理器执行所述计算机可读指令时还执行以下步骤：识别所述点云数据中各点与所述位置区域的位置关系；及当所述位置关系为点未在所述位置区域附近时，保留所述地面分割结果中未在所述位置区域附近的点对应的地面分割结果。
  15. 根据权利要求14所述的计算机设备，其特征在于，所述处理器执行所述计算机可读指令时还执行以下步骤：当所述位置关系为点位于所述位置区域附近时，计算位于所述位置区域附近的点与所述拟合平面之间的第一距离值；将所述第一距离值与第二阈值进行比较，根据比较结果确定所述位于所述位置区域附近的点对应的第一类别；及根据所述第一类别对所述地面分割结果进行更新，得到所述点云数据对应的目标地面分割结果。
  16. 一个或多个存储有计算机可读指令的非易失性计算机可读存储介质，所述计算机可读指令被一个或多个处理器执行时，使得所述一个或多个处理器执行以下步骤：
    获取点云数据；
    对所述点云数据进行地面分割，得到地面分割结果；
    对所述点云数据进行目标检测，得到所述点云数据对应的目标物体的三维包围框；
    提取所述三维包围框中的点云数据，在提取的点云数据中选取目标点云数据，将所述三维包围框对应的目标点云数据进行平面拟合，得到拟合平面；
    根据所述三维包围框提取各目标物体对应的位置区域；及
    根据所述位置区域以及所述拟合平面对所述地面分割结果进行更新，得到所述点云数据对应的目标地面分割结果。
  17. 根据权利要求16所述的存储介质，其特征在于，所述计算机可读指令被所述处理器执行时还执行以下步骤：根据预设参数将提取的点云数据划分为多个栅格；在各栅格对应的 点云数据中选取高度值最小的点；计算各栅格对应的点云数据中各点与所述高度值最小的点之间的高度差值；及选取高度差值小于第一阈值的点，根据选取的点得到目标点云数据。
  18. 根据权利要求17所述的存储介质，其特征在于，所述计算机可读指令被所述处理器执行时还执行以下步骤：所述三维包围框对应的多个栅格的目标点云数据进行组合，得到所述三维包围框对应的组合点云数据；及将所述三维包围框的组合点云数据进行平面拟合，得到所述三维包围框对应的拟合平面。
  19. 根据权利要求16所述的存储介质，其特征在于，所述计算机可读指令被所述处理器执行时还执行以下步骤：识别所述点云数据中各点与所述位置区域的位置关系；及当所述位置关系为点未在所述位置区域附近时，保留所述地面分割结果中未在所述位置区域附近的点对应的地面分割结果。
  20. 根据权利要求19所述的存储介质，其特征在于，所述计算机可读指令被所述处理器执行时还执行以下步骤：当所述位置关系为点位于所述位置区域附近时，计算位于所述位置区域附近的点与所述拟合平面之间的第一距离值；将所述第一距离值与第二阈值进行比较，根据比较结果确定所述位于所述位置区域附近的点对应的第一类别；及根据所述第一类别对所述地面分割结果进行更新，得到所述点云数据对应的目标地面分割结果。
PCT/CN2020/128139 2020-11-11 2020-11-11 基于点云数据的地面分割方法、装置和计算机设备 WO2022099511A1 (zh)

Priority Applications (2)

Application Number Priority Date Filing Date Title
PCT/CN2020/128139 WO2022099511A1 (zh) 2020-11-11 2020-11-11 基于点云数据的地面分割方法、装置和计算机设备
CN202080093095.XA CN114981840A (zh) 2020-11-11 2020-11-11 基于点云数据的地面分割方法、装置和计算机设备

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/CN2020/128139 WO2022099511A1 (zh) 2020-11-11 2020-11-11 基于点云数据的地面分割方法、装置和计算机设备

Publications (1)

Publication Number Publication Date
WO2022099511A1 true WO2022099511A1 (zh) 2022-05-19

Family

ID=81601896

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2020/128139 WO2022099511A1 (zh) 2020-11-11 2020-11-11 基于点云数据的地面分割方法、装置和计算机设备

Country Status (2)

Country Link
CN (1) CN114981840A (zh)
WO (1) WO2022099511A1 (zh)

Cited By (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115508807A (zh) * 2022-11-16 2022-12-23 苏州一径科技有限公司 点云数据的处理方法、装置、电子设备及存储介质
CN115719354A (zh) * 2022-11-17 2023-02-28 同济大学 基于激光点云提取立杆的方法与装置
CN116452403A (zh) * 2023-06-16 2023-07-18 瀚博半导体(上海)有限公司 点云数据处理方法、装置、计算机设备及存储介质
CN116486099A (zh) * 2023-06-21 2023-07-25 广州市易鸿智能装备有限公司 一种点云数据的数据提取方法和装置
CN116533998A (zh) * 2023-07-04 2023-08-04 深圳海星智驾科技有限公司 车辆的自动驾驶方法、装置、设备、存储介质及车辆
CN116740361A (zh) * 2023-08-11 2023-09-12 之江实验室 一种点云分割方法、装置、存储介质及电子设备
CN117058314A (zh) * 2023-08-16 2023-11-14 广州葛洲坝建设工程有限公司 一种基于点云数据的现浇结构模板逆向建模方法
CN117115774A (zh) * 2023-10-23 2023-11-24 锐驰激光(深圳)有限公司 草坪边界的识别方法、装置、设备及存储介质
CN117455928A (zh) * 2023-12-25 2024-01-26 珠海市格努科技有限公司 无序抓取过程中抓取对象的分割方法、装置和电子设备

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115375713B (zh) * 2022-10-25 2023-03-24 杭州华橙软件技术有限公司 地面点云的分割方法、装置及计算机可读存储介质
CN116823832B (zh) * 2023-08-29 2023-12-12 武汉精一微仪器有限公司 一种基于三维点云的锡膏缺陷检测方法

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP3644276A1 (en) * 2018-10-26 2020-04-29 Volvo Car Corporation Methods and systems for the fast estimation of three-dimensional bounding boxes and drivable surfaces using lidar point clouds
CN111429520A (zh) * 2020-03-02 2020-07-17 广州视源电子科技股份有限公司 负障碍物检测方法、装置、终端设备和存储介质
CN111507982A (zh) * 2019-06-28 2020-08-07 浙江大学 一种基于深度学习的点云语义分割方法
CN111639682A (zh) * 2020-05-13 2020-09-08 北京三快在线科技有限公司 基于点云数据的地面分割方法及装置
CN111652926A (zh) * 2020-04-30 2020-09-11 南京理工大学 基于稀疏点云数据的实时三维目标检测方法

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP3644276A1 (en) * 2018-10-26 2020-04-29 Volvo Car Corporation Methods and systems for the fast estimation of three-dimensional bounding boxes and drivable surfaces using lidar point clouds
CN111507982A (zh) * 2019-06-28 2020-08-07 浙江大学 一种基于深度学习的点云语义分割方法
CN111429520A (zh) * 2020-03-02 2020-07-17 广州视源电子科技股份有限公司 负障碍物检测方法、装置、终端设备和存储介质
CN111652926A (zh) * 2020-04-30 2020-09-11 南京理工大学 基于稀疏点云数据的实时三维目标检测方法
CN111639682A (zh) * 2020-05-13 2020-09-08 北京三快在线科技有限公司 基于点云数据的地面分割方法及装置

Cited By (16)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115508807A (zh) * 2022-11-16 2022-12-23 苏州一径科技有限公司 点云数据的处理方法、装置、电子设备及存储介质
CN115719354A (zh) * 2022-11-17 2023-02-28 同济大学 基于激光点云提取立杆的方法与装置
CN115719354B (zh) * 2022-11-17 2024-03-22 同济大学 基于激光点云提取立杆的方法与装置
CN116452403A (zh) * 2023-06-16 2023-07-18 瀚博半导体(上海)有限公司 点云数据处理方法、装置、计算机设备及存储介质
CN116452403B (zh) * 2023-06-16 2023-09-01 瀚博半导体(上海)有限公司 点云数据处理方法、装置、计算机设备及存储介质
CN116486099A (zh) * 2023-06-21 2023-07-25 广州市易鸿智能装备有限公司 一种点云数据的数据提取方法和装置
CN116533998A (zh) * 2023-07-04 2023-08-04 深圳海星智驾科技有限公司 车辆的自动驾驶方法、装置、设备、存储介质及车辆
CN116533998B (zh) * 2023-07-04 2023-09-29 深圳海星智驾科技有限公司 车辆的自动驾驶方法、装置、设备、存储介质及车辆
CN116740361B (zh) * 2023-08-11 2023-10-13 之江实验室 一种点云分割方法、装置、存储介质及电子设备
CN116740361A (zh) * 2023-08-11 2023-09-12 之江实验室 一种点云分割方法、装置、存储介质及电子设备
CN117058314A (zh) * 2023-08-16 2023-11-14 广州葛洲坝建设工程有限公司 一种基于点云数据的现浇结构模板逆向建模方法
CN117058314B (zh) * 2023-08-16 2024-04-12 广州葛洲坝建设工程有限公司 一种基于点云数据的现浇结构模板逆向建模方法
CN117115774A (zh) * 2023-10-23 2023-11-24 锐驰激光(深圳)有限公司 草坪边界的识别方法、装置、设备及存储介质
CN117115774B (zh) * 2023-10-23 2024-03-15 锐驰激光(深圳)有限公司 草坪边界的识别方法、装置、设备及存储介质
CN117455928A (zh) * 2023-12-25 2024-01-26 珠海市格努科技有限公司 无序抓取过程中抓取对象的分割方法、装置和电子设备
CN117455928B (zh) * 2023-12-25 2024-04-02 珠海市格努科技有限公司 无序抓取过程中抓取对象的分割方法、装置和电子设备

Also Published As

Publication number Publication date
CN114981840A (zh) 2022-08-30

Similar Documents

Publication Publication Date Title
WO2022099511A1 (zh) 基于点云数据的地面分割方法、装置和计算机设备
CN111160302B (zh) 基于自动驾驶环境的障碍物信息识别方法和装置
CN110458854B (zh) 一种道路边缘检测方法和装置
WO2022099530A1 (zh) 点云数据运动分割方法、装置、计算机设备和存储介质
WO2018068653A1 (zh) 点云数据处理方法、装置及存储介质
WO2022016311A1 (zh) 基于点云的三维重建方法、装置和计算机设备
CN111179274B (zh) 地图地面分割方法、装置、计算机设备和存储介质
CN110197173B (zh) 一种基于双目视觉的路沿检测方法
WO2021134285A1 (zh) 图像跟踪处理方法、装置、计算机设备和存储介质
CN113008296B (zh) 用于通过在点云平面上融合传感器数据来检测汽车环境的方法和汽车控制单元
JP2019191991A (ja) 点群から対象の向きを推定する対象情報推定装置、プログラム及び方法
CN112001378B (zh) 基于特征空间的车道线处理方法、装置、车载终端和介质
WO2022226831A1 (zh) 未定义类别障碍物的检测方法、装置和计算机设备
CN115439621A (zh) 一种煤矿井下巡检机器人三维地图重建及目标检测方法
CN109115232B (zh) 导航的方法和装置
CN112986982B (zh) 环境地图参照定位方法、装置和移动机器人
WO2022078342A1 (zh) 动态占据栅格估计方法及装置
US11415698B2 (en) Point group data processing device, point group data processing method, point group data processing program, vehicle control device, and vehicle
CN113313629B (zh) 交叉路口自动识别方法、系统及其模型保存方法、系统
CN111476879A (zh) 一种点云数据的处理方法、终端及存储介质
CN116052120A (zh) 基于图像增强和多传感器融合的挖掘机夜间物体检测方法
US20220171061A1 (en) Apparatus and method for recognizing high-elevation structure using lidar sensor
Mattson et al. Reducing ego vehicle energy-use by LiDAR-based lane-level positioning
CN117523522B (zh) 隧道场景的识别方法、装置、设备和存储介质
CN111257903B (zh) 车辆定位方法、装置、计算机设备和存储介质

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 20961070

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 20961070

Country of ref document: EP

Kind code of ref document: A1