Calibration method and device of unmanned container truck positioning system, electronic equipment and readable storage medium

Abstract

The invention discloses a calibration method and device of an unmanned container truck positioning system, electronic equipment and a readable storage medium, and relates to the technical field of automatic driving. The method comprises the following steps: obtaining indication information; obtaining M frames of point cloud data; obtaining a first type of effective points and a second type of effective points corresponding to the operation item from each frame of point cloud data, if the operation item is a container unloading item, the first type of effective points belong to points generated on the front end face of the container, and the second type of effective points belong to points generated on the side face of the container, and if the operation item is a container loading item, the second type of effective points belong to points generated on the side face of the container; the first type of effective points belong to points generated at the front end of the vehicle head, and the second type of effective points belong to points generated on the side face of the chassis; and determining an X-axis calibration value of the work item based on the X-axis coordinate values of the first type of effective points, and determining a Z-axis calibration value of the work item based on the Z-axis coordinate values of the second type of effective points. According to the technical scheme disclosed by the invention, the operation position of the operation item can be accurately calibrated, so that the operation efficiency of the unmanned container truck is improved.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of automatic driving, and particularly relates to a calibration method and device of an unmanned container truck positioning system, an electronic device and a readable storage medium. BACKGROUND

[0002] Unmanned container trucks (referred to as unmanned container trucks) are widely used in logistics transportation scenes such as ports and warehouses. A container-vehicle positioning system (CPS) plays an important role in the automatic operation process of the unmanned container truck.

[0003] The working principle of the unmanned container truck positioning system is as follows: the position information of the unmanned container truck is collected by using a sensor, the position information of the unmanned container truck is compared with a pre-calibrated operation position (a position corresponding to a loading and unloading device, which needs to be pre-calibrated), so as to control the position adjustment of the unmanned container truck, and in the case that the unmanned container truck reaches the operation position, it is indicated that the unmanned container truck and the loading and unloading device are aligned, and a positioning signal is sent to the loading and unloading device, so that the loading and unloading device can perform loading and unloading operations. Based on the unmanned container truck positioning system, the position of the unmanned container truck can be adjusted in time, so as to reduce the time consumed by the unmanned container truck and the loading and unloading device for alignment, and to improve the operation efficiency of the unmanned container truck.

[0004] It can be seen that the unmanned container truck positioning system controls the position adjustment of the unmanned container truck based on the position information of the unmanned container truck and the pre-calibrated operation position, and whether the pre-calibrated operation position is accurate directly relates to the operation efficiency of the unmanned container truck. For those skilled in the art, how to accurately calibrate the operation position corresponding to the operation item in the unmanned container truck system is a problem to be solved. SUMMARY

[0005] In view of the above problems, the application provides a calibration method and device of an unmanned container truck positioning system, an electronic device and a readable storage medium, so as to accurately calibrate the operation position corresponding to the operation item in the unmanned container truck positioning system.

[0006] To achieve the above-mentioned purpose, the application provides the following technical solutions:

[0007] In a first aspect, the application provides a calibration method of an unmanned container truck positioning system, comprising:

[0008] obtaining indication information, the indication information comprising an operation item and an operation lane;

[0009] obtaining M frames of point cloud data, the M frames of point cloud data being collected by a multi-line laser radar when the unmanned container truck stops at an expected position of the operation item in the operation lane, M being greater than or equal to 1;

[0010] obtain a first type of valid point corresponding to the job item from each frame of point cloud data, wherein in a case where the job item is an unloading box job, the first type of valid point belongs to a point generated by a front end face of a container carried by the unmanned truck, and in a case where the job item is a loading box job, the first type of valid point belongs to a point generated by a front end of a truck head of the unmanned truck;

[0011] determine an X-axis calibration value of the job item based on an X-axis coordinate value of the first type of valid point in the M frames of point cloud data;

[0012] obtain a second type of valid point corresponding to the job item from each frame of point cloud data, wherein in a case where the job item is an unloading box job, the second type of valid point belongs to a point generated by a side face of the container carried by the unmanned truck, and in a case where the job item is a loading box job, the second type of valid point belongs to a point generated by a side face of a chassis of the unmanned truck;

[0013] determine a Z-axis calibration value of the job item based on a Z-axis coordinate value of the second type of valid point in the M frames of point cloud data;

[0014] wherein the X-axis is a length direction of a lane, and the Z-axis is a width direction of the lane.

[0015] In a second aspect, the present application provides a calibration device of an unmanned truck positioning system, comprising:

[0016] an indication information obtaining module, configured to obtain indication information, the indication information comprising a job item and a job lane;

[0017] a point cloud data obtaining module, configured to obtain M frames of point cloud data, the M frames of point cloud data being collected by a multi-line laser radar in a case where the unmanned truck is parked at an expected position of the job item in the job lane, M being greater than or equal to 1;

[0018] a first type of valid point obtaining module, configured to obtain a first type of valid point corresponding to the job item from each frame of point cloud data, wherein in a case where the job item is an unloading box job, the first type of valid point belongs to a point generated by a front end face of a container carried by the unmanned truck, and in a case where the job item is a loading box job, the first type of valid point belongs to a point generated by a front end of a truck head of the unmanned truck;

[0019] a first calibration module, configured to determine an X-axis calibration value of the job item based on an X-axis coordinate value of the first type of valid point in the M frames of point cloud data;

[0020] the second type of valid points belong to points generated by the side of the container carried by the unmanned truck, in the case that the work item is a loading item, the second type of valid points belong to points generated by the side of the chassis of the unmanned truck;

[0021] the second calibration module is configured to determine a Z-axis calibration value of the work item based on the Z-axis coordinate values of the second type of valid points in the M frames of point cloud data;

[0022] wherein, the X-axis is the length direction of the lane, and the Z-axis is the width direction of the lane.

[0023] In a third aspect, the present application provides an electronic device, comprising at least one processor and a memory connected to the processor, wherein:

[0024] the memory is configured to store a computer program;

[0025] the processor is configured to execute the computer program, so that the electronic device implements the calibration method described above.

[0026] In a fourth aspect, the present application provides a readable storage medium, the readable storage medium carries one or more computer programs, when the one or more computer programs are executed by an electronic device, so that the electronic device implements the calibration method described above.

[0027] Therefore, the beneficial effects of the present application are:

[0028] The technical solution disclosed in the present application can accurately calibrate the work position corresponding to the work item, thereby improving the work efficiency of the unmanned truck. BRIEF DESCRIPTION OF DRAWINGS

[0029] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed to be used in the embodiments or prior art description will be briefly introduced. Obviously, the drawings in the following description are some embodiments of the present application, and those skilled in the art can also obtain other drawings according to these drawings without creative labor.

[0030] Figure 1A flowchart of a calibration method of an unmanned container truck positioning system disclosed in the present application;

[0031] Figure 2 A working scene diagram of an unmanned container truck disclosed in the present application;

[0032] Figure 3 A structural diagram of a calibration device of an unmanned container truck positioning system disclosed in the present application;

[0033] Figure 4 A structural diagram of an electronic device disclosed in the present application. DETAILED DESCRIPTION

[0034] The present application discloses a calibration method, device, electronic device and readable storage medium of an unmanned container truck positioning system, to accurately calibrate the working position corresponding to the working item in the unmanned container truck positioning system.

[0035] To make the objectives, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are some but not all of the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the protection scope of the present application.

[0036] To facilitate the understanding of the technical solutions disclosed in the present application, the working site and working process of the unmanned container truck are described. The working site is arranged with loading and unloading equipment, and the working site is usually provided with multiple lanes. The unmanned container truck travels along the designated lane to the pre-calibrated working position, i.e. completes the alignment with the loading and unloading equipment, and the loading and unloading equipment unloads the container carried by the unmanned container truck or places the container on the unmanned container truck, and then the unmanned container truck leaves the working position. The type of the loading and unloading equipment is various, for example, it can be a shore-based bridge device or a yard bridge device.

[0037] In the present solution, the sensor in the unmanned container truck positioning system adopts a multi-line laser radar. Here, the multi-line laser radar is introduced:

[0038] The multi-line laser radar has multiple laser transceiver modules, wherein each laser transceiver module includes a laser transmitter and a laser receiver. For example, a 4-line laser radar has 4 laser transceiver modules, an 8-line laser radar has 8 laser transceiver modules, and a 16-line laser radar has 16 laser transceiver modules. The higher the line number of the multi-line laser radar, i.e. the more the number of laser transceiver modules, the smaller the vertical resolution of the multi-line laser radar, and the more perfect the obtained object surface profile.

[0039] In the working process of the multi-line laser radar, a plurality of laser transceiver modules are driven by a driving mechanism to rotate in the horizontal direction, and each laser transceiver module collects a plurality of points. In a scanning period, the number of points collected by each laser transceiver module is related to the scanning frequency. For example, if the horizontal resolution of the multi-line laser radar is 0.2°, then the number of points collected by each laser transceiver module in a scanning period is: the angle rotated in a scanning period / 0.2°. For example, if the horizontal resolution of the multi-line laser radar is 0.4°, then the number of points collected by each laser transceiver module in a scanning period is: the angle rotated in a scanning period / 0.4°.

[0040] The points collected by the multi-line laser radar in a scanning period form a frame of point cloud data. Taking a 64-line laser radar as an example, assuming that the horizontal resolution of the 64-line laser radar is 0.2, and the driving mechanism drives the laser transceiver module to rotate 360° in a scanning period, then the maximum number of points collected by the 64-line laser radar in a scanning period is: (360° / 0.2°)*64, i.e. 115200, and these points form a frame of point cloud data. The data of each point in the point cloud data includes: three-dimensional coordinate values and laser reflection intensity.

[0041] The installation position of the multi-line laser radar can be various. First, the multi-line laser radar is located above the middle of the lane, which can be understood as that the distance between the projection of the multi-line laser radar on the ground and the center line of the lane is less than a preset limit value; second, the multi-line laser radar is located on one side of the lane. From the detection effect, only for the two sides of the unmanned straddle carrier and the container, the first mode is adopted to arrange the multi-line laser radar, and in the process of the unmanned straddle carrier moving towards the loading and unloading equipment, the multi-line laser radar can detect the two sides of the unmanned straddle carrier and the container; the second mode is adopted to arrange the multi-line laser radar, and in the process of the unmanned straddle carrier moving towards the loading and unloading equipment, the multi-line laser radar can only detect the side of the unmanned straddle carrier and the container close to the multi-line laser radar.

[0042] Referring to Figure 1 , Figure 1 A flowchart of a calibration method of an unmanned straddle carrier positioning system disclosed in the present application. The method is executed by an electronic device, comprising:

[0043] S10: obtaining indication information.

[0044] The indication information includes a work item and a work lane.

[0045] The indication information is input by the staff according to the needs, and the indication information contains the work item to be calibrated and the work lane where the unmanned straddle carrier is located. The work item includes a container loading item and a container unloading item. Optionally, the container unloading item is further divided into pre-unloading small containers, post-unloading small containers, double-unloading containers, and single large-unloading containers.

[0046] S20: Obtain M frames of point cloud data.

[0047] The M frames of point cloud data are collected by the multi-line laser radar when the unmanned truck is parked at the desired position of the work project in the work lane, and M is greater than or equal to 1.

[0048] The worker controls the unmanned truck to park at the desired position of the current work project. When the unmanned truck is parked at the desired position of the current work project, scanning is performed by the multi-line laser radar to generate point cloud data. In step S20, M frames of point cloud data generated by the multi-line laser radar are obtained, and the calibration of the work position of the current work project is completed based on the M frames of point cloud data. The value of M can be 1 or an integer greater than 1.

[0049] S30: Obtain the first type of valid points corresponding to the work project from each frame of point cloud data.

[0050] In the case of an unloading project, the first type of valid points belong to the points generated by the front end of the container carried by the unmanned truck, and in the case of a loading project, the first type of valid points belong to the points generated by the front end of the truck head of the unmanned truck.

[0051] S40: Determine the X-axis calibration value of the work project based on the X-axis coordinate value of the first type of valid points in the M frames of point cloud data.

[0052] S50: Obtain the second type of valid points corresponding to the work project from each frame of point cloud data.

[0053] In the case of an unloading project, the second type of valid points belong to the points generated by the side of the container carried by the unmanned truck, and in the case of a loading project, the second type of valid points belong to the points generated by the side of the chassis of the unmanned truck.

[0054] The structure of the unmanned truck is described here. The unmanned truck includes a truck head and a chassis (which can also be referred to as a tray), and the chassis is used to carry containers. Optionally, the unmanned truck can also include a truck tail. As shown in Figure 2 The unmanned truck shown in FIG. 1 includes a truck head, a truck tail, and a chassis, wherein the truck head and the truck tail are each provided with a power device, and the structures of the truck head and the truck tail are similar, with the truck head being referred to as the truck head in the forward direction and the truck tail being referred to as the truck tail in the forward direction.

[0055] S60: Determine the Z-axis calibration value of the work project based on the Z-axis coordinate value of the second type of valid points in the M frames of point cloud data.

[0056] In the scheme, a three-dimensional coordinate system is constructed with the multi-line laser radar as the origin, the X-axis of the three-dimensional coordinate system is the length direction of the lane, the Y-axis is the vertical direction, and the Z-axis is the width direction of the lane.

[0057] In the case where the work item is the unloading box item, the unmanned container truck carries the container, and the point cloud data generated by the multi-line laser radar includes points generated by the ground and points generated by the unmanned container truck and the container. The first type of valid points corresponding to the unloading box item are obtained from the point cloud data, and the first type of valid points belong to the points generated by the front end face of the container carried by the unmanned container truck. The second type of valid points corresponding to the unloading box item are obtained from the point cloud data, and the second type of valid points belong to the points generated by the side face of the container carried by the unmanned container truck.

[0058] Optionally, in the case where the work item is the unloading box item, the first type of valid points obtained from each frame of point cloud data are all the points generated by the front end face of the container, or are a part of the points generated by the front end face of the container. Optionally, in the case where the work item is the unloading box item, the second type of valid points obtained from each frame of point cloud data are all the points generated by the side face of the container, or are a part of the points generated by the side face of the container.

[0059] In the case where the work item is the loading box item, the unmanned container truck does not carry the container, and the point cloud data generated by the multi-line laser radar includes points generated by the ground and points generated by the unmanned container truck. Specifically, in the case where the unmanned container truck is not provided with a trailer, the points generated by the unmanned container truck are points generated by the front end of the vehicle head and the chassis of the unmanned container truck, and in the case where the unmanned container truck is provided with a trailer, the points generated by the unmanned container truck further include points generated by the trailer. The first type of valid points corresponding to the loading box item are obtained from the point cloud data, and the first type of valid points belong to the points generated by the front end of the vehicle head of the unmanned container truck. The second type of valid points corresponding to the loading box item are obtained from the point cloud data, and the second type of valid points belong to the points generated by the side face of the chassis of the unmanned container truck.

[0060] Optionally, in the case where the work item is the loading box item, the first type of valid points obtained from each frame of point cloud data are all the points generated by the front end of the vehicle head of the unmanned container truck, or are a part of the points generated by the front end of the vehicle head of the unmanned container truck. Optionally, in the case where the work item is the loading box item, the second type of valid points obtained from each frame of point cloud data are all the points generated by the side face of the chassis of the unmanned container truck, or are a part of the points generated by the side face of the chassis of the unmanned container truck.

[0061] After the first type of valid points corresponding to the work item are obtained from M frames of point cloud data, the X-axis calibration value of the work item is determined based on the X-axis coordinate values of the first type of valid points. After the second type of valid points corresponding to the work item are obtained from M frames of point cloud data, the Z-axis calibration value of the work item is determined based on the Z-axis coordinate values of the second type of valid points.

[0062] The calibration method of the unmanned container truck positioning system disclosed in the application is as follows: when the unmanned container truck stops at the expected position of the work project in the work lane, scanning is performed by the multi-line laser radar; M frames of point cloud data collected by the multi-line laser radar are obtained; first type effective points and second type effective points are obtained from each frame of point cloud data, wherein, when the work project is a container unloading project, the first type effective points are points generated by the front end face of the container carried by the unmanned container truck, and the second type effective points are points generated by the side face of the container carried by the unmanned container truck, when the work project is a container loading project, the first type effective points are points generated by the front end of the truck head of the unmanned container truck, and the second type effective points are points generated by the side face of the chassis of the unmanned container truck; the X-axis calibration value of the work project is determined based on the X-axis coordinate value of the first type effective points in the M frames of point cloud data, and the Z-axis calibration value of the work project is determined based on the Z-axis coordinate value of the second type effective points in the M frames of point cloud data. The technical solution disclosed in the application can stop the unmanned container truck at the corresponding expected position in the work lane for different work projects, generate point cloud data by scanning with the multi-line laser radar, then obtain the first type effective points and the second type effective points corresponding to the work project from the point cloud data, and then determine the X-axis calibration value of the work project based on the X-axis coordinate value of the first type effective points and determine the Z-axis calibration value of the work project based on the Z-axis coordinate value of the second type effective points, so that the work position corresponding to the work project can be accurately calibrated, thereby improving the work efficiency of the unmanned container truck.

[0063] Please refer to Figure 2 , Figure 2 for the schematic diagram of the work scene of the unmanned container truck. In Figure 2 , 100 is a loading and unloading device, 200 is an unmanned container truck, and 300 is a container. In Figure 2 , in the three-dimensional coordinate system constructed with the multi-line laser radar as the origin, the X-axis is the length direction of the lane, the Y-axis is the vertical direction, and the Z-axis is the width direction of the lane. Specifically, the positive direction of the X-axis is the opposite direction of the forward direction of the unmanned container truck, that is, the negative direction of the X-axis is the forward direction of the unmanned container truck, the positive direction of the Y-axis is the vertical upward direction, and the positive direction of the Z-axis is the left side of the lane. In the following, the three-dimensional coordinate system shown in Figure 2 will be taken as an example to make a more detailed description of the scheme disclosed in the application. Based on the three-dimensional coordinate system shown in Figure 2 , the X-axis coordinate value of the point of the truck head of the unmanned container truck is less than the X-axis coordinate value of the point of the truck tail of the unmanned container truck.

[0064] The following will focus on how to obtain the first type effective points corresponding to the work project from the point cloud data.

[0065] In another embodiment of the present application, in the case of the job item being a box unloading item, the first type of valid points corresponding to the box unloading item are obtained from any frame of point cloud data, and the following scheme is adopted:

[0066] A1: Obtain a first point set generated by the container from the point cloud data.

[0067] The Z-axis coordinate value interval corresponding to each work lane is calibrated in advance, that is, the Z-axis coordinate value interval of the points generated by each work lane, the unmanned container truck located in the work lane, and the container carried by the unmanned container truck (in the case of a box unloading item) is known.

[0068] Optionally, the first point set is obtained from the point cloud data, and the following scheme is adopted:

[0069] 1) According to the Z-axis coordinate value interval of the current work lane, obtain the point cloud data generated by the current work lane, the unmanned container truck located in the work lane, and the container carried by the unmanned container truck from the point cloud data.

[0070] 2) The height of the front of the unmanned container truck is a known quantity, and the point cloud data generated by the container is obtained from the point cloud data obtained in the previous step based on the height of the front of the unmanned container truck. For ease of description, the set of points generated by the container is referred to as the first point set. It can be understood that the height of the front of the unmanned container truck is a known quantity, and a part of the container may be blocked by the front, so the point cloud data obtained in the previous step is processed based on the height of the front of the unmanned container truck to filter out the points generated by the road surface and the unmanned container truck, and retain the points generated by the container.

[0071] Here is an example: assume that the multi-line laser radar is installed at a distance of 12.6 meters from the ground, and the height of the front of the unmanned container truck is 1.6 meters (i.e., the distance between the highest point in the front and the ground is 1.6 meters), and accordingly, the highest point of the front is at Figure 2 The Y-axis coordinate value of the three-dimensional coordinate system shown in the figure is -11, and the points with a Y-axis coordinate value greater than the first value are obtained from the point cloud data obtained in the previous step, wherein the first value is greater than -11, and the obtained points are generated by the container. It can be understood that setting the first value to a larger value can reduce the possibility of misjudging the points generated by the unmanned container truck as the points generated by the container.

[0072] As an optional scheme, the value of the first value is adjusted so that only the points generated by the upper half of the unmanned container truck are retained as the first point set.

[0073] A2: Determine the first minimum value in the X-axis coordinate values of the points in the first point set.

[0074] The X-axis coordinate values of the points in the first point set are compared to determine a minimum value in the X-axis coordinate values of the points, which is referred to as a first minimum value for convenience of description.

[0075] A3: For each channel of the multi-line laser radar, a point with the minimum X-axis coordinate value among the points belonging to the channel in the first point set is determined as a first-type candidate point.

[0076] As introduced in the foregoing, the multi-line laser radar includes a plurality of laser transceiver modules, and each laser transceiver module collects a plurality of points in one scanning period. Each laser transceiver module is regarded as a channel, and the points belonging to a certain channel refer to the points collected by the channel (i.e., the laser transceiver module).

[0077] In this step, for each channel of the multi-line laser radar, a point with the minimum X-axis coordinate value among the points belonging to the channel in the first point set is determined as a first-type candidate point.

[0078] It should be noted that the first point set generated by the container is part of the original point cloud data, and the number of channels to which the points in the first point set belong is less than the number of channels (i.e., the number of laser transceiver modules) of the multi-line laser radar. That is, the first point set can not include points generated by certain channels, and accordingly, the first-type candidate point obtained for the channel is empty.

[0079] Taking a 64-line laser radar used in an unmanned container truck positioning system as an example, the points in the point cloud data generated by the 64-line laser radar belong to 64 channels. The first point set can not include points belonging to a certain channel or certain channels. For the first channel of the 64-line laser radar, a point with the minimum X-axis coordinate value among the points belonging to the first channel in the first point set is determined as a first-type candidate point; for the second channel of the 64-line laser radar, a point with the minimum X-axis coordinate value among the points belonging to the second channel in the first point set is determined as a first-type candidate point; and the processing of other channels is similar, which is not described herein.

[0080] A4: The difference between the X-axis coordinate value of each first-type candidate point and the first minimum value is determined, and the first-type candidate point with a difference less than a first threshold value is determined as a first-type valid point.

[0081] For each first-type candidate point, the difference between the X-axis coordinate value of the first-type candidate point and the first minimum value is determined, if the difference is less than the first threshold value, the first-type candidate point is taken as a first-type valid point, and if the difference is greater than or equal to the first threshold value, the first-type candidate point is discarded. Optionally, the first threshold value is set to 10 cm.

[0082] Correspondingly, the X-axis calibration value of the unloading project is determined based on the X-axis coordinate values of the first type of effective points in the M-frame point cloud data, and the following scheme is adopted: the average value of the X-axis coordinate values of the first type of effective points in the M-frame point cloud data is determined as the X-axis calibration value of the unloading project. That is, in the case of the work project being the unloading project, the X-axis calibration value of the work project is determined based on the X-axis coordinate values of the first type of effective points in the M-frame point cloud data, which includes: determining the average value of the X-axis coordinate values of the first type of effective points in the M-frame point cloud data as the X-axis calibration value of the work project.

[0083] In the above embodiments of the present application, in the case of the work project being the unloading project, the scheme for determining the first type of effective points and the scheme for determining the X-axis calibration value of the unloading project based on the first type of effective points have the following technical advantages:

[0084] 1. High precision. By selecting the minimum value (i.e., the first minimum value) of the X-axis coordinate values of the points in the first point set, for each channel of the multi-line laser radar, selecting the point with the smallest X-axis coordinate value among the points belonging to that channel in the first point set as the first type of candidate point, and then filtering according to the difference between the X-axis coordinate value of the first type of candidate point and the first minimum value, it can be ensured that the point closest to the leading edge of the container is found in the point cloud data as the first type of effective point. This method can more accurately extract effective information from a large amount of point cloud data, reducing interference from noise or irrelevant data (such as the ground and unmanned trucks).

[0085] 2. Improve automation and robustness. By automatically determining the minimum X-axis coordinate point of each channel and comparing its difference with the first minimum value, manual intervention can be reduced, and the robustness of the system can be improved. In particular, when the quality of the point cloud data is not high or the environmental conditions are complex, this automated processing method can still ensure relatively stable and reliable results.

[0086] 3. Improve flexibility and adaptability. By setting a first threshold (e.g., 10 cm) to determine the selection criteria for the first type of effective points, the size of the first threshold can be adjusted according to actual needs to adapt to different environments or different precision requirements in application scenarios. For example, in some special scenarios, higher precision data may be required to ensure accurate positioning and unloading operations, and the accuracy of selection can be improved by adjusting the size of the first threshold.

[0087] 4. Help improve calibration accuracy. By calculating the average value of the X-axis coordinate values of the first type of effective points as the X-axis calibration value of the unloading project, it helps to improve the accuracy of the calibration process. This way, the most representative effective points in the point cloud data are used as a reference to obtain more reliable calibration results, thereby improving the positioning and control accuracy of the unmanned truck.

[0088] 5. Suitable for different types of laser radar. Even if the number of channels of the laser radar is different, this scheme is still applicable. For example, 64-line laser radar and other line number laser radars may be different in data processing, but through this scheme, the first type of valid points can be independently screened for each channel, regardless of the number of radar lines, and has good versatility.

[0089] In another embodiment of the present application, in the case of a loading project, the first type of valid points corresponding to the loading project are obtained from any frame of point cloud data, and the following scheme is adopted:

[0090] B1: Obtain a second point set from the point cloud data, and the second point set contains points generated by the front of the unmanned truck.

[0091] Optionally, the second point set is obtained from the point cloud data, and the following scheme is adopted:

[0092] 1) According to the Z-axis coordinate value interval of the current work lane, the point cloud data generated by the unmanned truck in the current work lane is obtained from the point cloud data.

[0093] 2) From the point cloud data obtained in the previous step, the points generated by the ground and the two side surfaces of the chassis of the unmanned truck are filtered out to obtain a second point set, and the second point set contains points generated by the front of the unmanned truck. In the case where the unmanned truck is provided with a tail, the second point set also contains points generated by the tail of the unmanned truck. In addition, the second point set may also contain some impurity points.

[0094] Here is an example.

[0095] First, according to the Z-axis coordinate value interval of the current work lane, the point cloud data whose Z-axis coordinate value is located in the Z-axis coordinate value interval is obtained from the point cloud data, that is, the point cloud data generated by the unmanned truck in the current work lane is obtained.

[0096] Second, for the point cloud data obtained in the previous step, according to the Y-axis coordinate value of each point and the pre-constructed Y-axis first screening interval, the point cloud data is processed, and only the points whose Y-axis coordinate value is located in the Y-axis first screening interval are retained, that is, the point cloud data generated by the ground is filtered out, and the point cloud data generated by the unmanned truck is retained.

[0097] Among them, the Y-axis first screening interval is determined according to the distance between the laser radar and the ground, the chassis height of the unmanned truck, and the height of the front of the unmanned truck. The chassis height refers to the minimum ground clearance of the chassis.

[0098] It can be understood that if the lower limit value of the Y-axis first screening interval is set to a larger value, the possibility of the points generated by the ground being mistakenly retained can be reduced, on the other hand, if the lower limit value of the Y-axis first screening interval is too large, part of the points generated by the unmanned truck may be filtered out. Based on the foregoing considerations, as an optional solution, the lower limit value of the Y-axis first screening interval is determined according to the Y-axis coordinate value of the lowest point of the chassis of the unmanned truck. In addition, if the upper limit value of the Y-axis first screening interval is set to a smaller value, part of the points generated by the unmanned truck may be filtered out. Based on the foregoing considerations, as an optional solution, the upper limit value of the Y-axis first screening interval is determined according to the Y-axis coordinate value of the highest point of the front of the unmanned truck.

[0099] For example, the multi-line laser radar is installed at a position 12.6 meters away from the ground, the height of the chassis of the unmanned truck is 1.1 meters, and the height of the front of the unmanned truck is 1.6 meters. The Y-axis first screening interval can be set to [-11.5, -10.5], and for each point in the point cloud data, the points with Y-axis coordinate values within the Y-axis first screening interval are retained, and other points are filtered out, thereby obtaining points generated by the unmanned truck.

[0100] Third, for the point cloud data generated by the unmanned truck obtained in the previous step, the maximum and minimum values of the Z-axis coordinate values of each point are determined, the Z-axis first screening interval is determined based on the maximum value (the upper limit value and the lower limit value of the Z-axis first screening interval are: the maximum value plus a first correction value, the maximum value minus a first correction value. For example, the first correction value is 10 cm), the Z-axis second screening interval is determined based on the minimum value (the upper limit value and the lower limit value of the Z-axis second screening interval are: the minimum value plus a second correction value, the minimum value minus a second correction value. For example, the second correction value is 10 cm), and for the point cloud data generated by the unmanned truck, the points with Z-axis coordinate values within the Z-axis first screening interval and the Z-axis second screening interval are filtered out, thereby filtering out the points generated by the two sides of the chassis of the unmanned truck. It should be noted that the chassis of the unmanned truck for carrying containers is usually a frame structure, and the points generated by the part of the chassis in contact with the container are less and can be ignored.

[0101] In the above scheme, the point cloud data generated by the multi-line laser radar for the current working lane is obtained first, then the points generated by the unmanned truck are obtained based on the Y-axis coordinate values of each point and the Y-axis first screening interval, and then the points generated by the two sides of the chassis of the unmanned truck are filtered out based on the Z-axis coordinate values of each point.

[0102] The above scheme can be replaced by:

[0103] First, for the point cloud data generated by the multi-line laser radar, according to the Y-axis coordinate value of each point and the pre-constructed Y-axis first screening interval, the point cloud data is processed to filter out the points generated by the ground in the point cloud data.

[0104] Second, for the point cloud data obtained in the previous step, according to the Z-axis coordinate value interval of the current working lane, the point cloud data whose Z-axis coordinate value is within the Z-axis coordinate value range is obtained from the point cloud data, that is, the points generated by the unmanned container truck in the current working lane are obtained.

[0105] Third, for each point obtained in the previous step, the maximum and minimum values of the Z-axis coordinate value of each point are determined, the Z-axis first screening interval is determined based on the maximum value (the upper limit value and the lower limit value of the Z-axis first screening interval are: the maximum value plus a first correction value, and the maximum value minus a first correction value. For example, the first correction value is 10 cm), and the Z-axis second screening interval is determined based on the minimum value (the upper limit value and the lower limit value of the Z-axis second screening interval are: the minimum value plus a second correction value, and the minimum value minus a second correction value. For example, the second correction value is 10 cm), and the points whose Z-axis coordinate value is within the Z-axis first screening interval and the Z-axis second screening interval are filtered out, thereby filtering out the points generated by the two sides of the chassis of the unmanned container truck in the point cloud data.

[0106] B2: For each channel of the multi-line laser radar, in the plurality of points belonging to the channel in the second point set, the point with the minimum X-axis coordinate value is determined as the second type of candidate point.

[0107] In this step, for each channel of the laser radar, in the points generated by the channel and contained in the second point set, the point with the minimum X-axis coordinate value is determined as the second type of candidate point.

[0108] It should be noted that the second point set generated by the front of the unmanned container truck is part of the original point cloud data, and the number of channels to which the points in the second point set belong is less than the number of channels of the multi-line laser radar. That is, the second point set may not contain points generated by some channels, and accordingly, the second type of candidate point obtained for the channel is empty.

[0109] Taking a 64-line laser radar used in an unmanned container truck positioning system as an example. The points in the point cloud data generated by the 64-line laser radar belong to 64 channels. The second point set may not contain points belonging to one or more channels. For the first channel of the 64-line laser radar, in the points generated by the first channel and contained in the second point set, the point with the minimum X-axis coordinate value is determined as the second type of candidate point; for the second channel of the 64-line laser radar, in the points generated by the second channel and contained in the second point set, the point with the minimum X-axis coordinate value is determined as the second type of candidate point; the processing of other channels is similar, which will not be described here.

[0110] B3: determining the boundary position of the truck head of the unmanned container truck based on the second type of candidate points.

[0111] Optionally, the boundary position of the truck head of the unmanned container truck is determined based on the second type of candidate points, and the following scheme is adopted:

[0112] First, the second type of candidate points are sorted according to the X-axis coordinate values.

[0113] Second, the minimum effective value is determined based on the X-axis coordinate values of the sorted second type of candidate points, wherein the minimum effective value is the minimum value of all X-axis coordinate values contained in all second type of candidate points whose adjacent positions after sorting and X-axis coordinate values have a difference less than a third threshold value.

[0114] Optionally, the second type of candidate points are sorted in ascending order of X-axis coordinate values, and in the sorted second type of candidate points, the second type of candidate point with the smallest X-axis coordinate value has the smallest serial number; i is assigned a value of 1; the difference between the X-axis coordinate values of the second type of candidate point located at the i-th position and the second type of candidate point located at the i+1-th position after sorting is calculated; if the difference is less than the third threshold value, the X-axis coordinate value of the second type of candidate point located at the i-th position is determined as the minimum effective value; if the difference is greater than or equal to the third threshold value, the value of i is increased by 1, and the process is repeated until the difference between the X-axis coordinate values of the second type of candidate point located at the i-th position and the second type of candidate point located at the i+1-th position is less than the third threshold value, and the X-axis coordinate value of the second type of candidate point located at the i-th position is determined as the minimum effective value. Optionally, the third threshold value is set to 10 cm.

[0115] Optionally, the second type of candidate points are sorted in ascending order of X-axis coordinate values, and in the sorted second type of candidate points, the second type of candidate point with the smallest X-axis coordinate value has the smallest serial number; for each pair of adjacent second type of candidate points, the difference between the X-axis coordinate values of the two second type of candidate points is calculated; the second type of candidate points with a difference less than the third threshold value are screened out, and the second type of candidate point with the smallest serial number is determined from all the screened second type of candidate points, and the X-axis coordinate value of the second type of candidate point is determined as the minimum effective value.

[0116] The above scheme for determining the minimum effective value is a filtering method based on local similarity, which has the following advantages:

[0117] 1. Remove noise and outliers: In actual point cloud data or sensor data, there may be some noise or outliers due to device errors, environmental influences, etc. The difference between adjacent points of most effective data is usually small, so by calculating the difference between adjacent points and filtering out points with large differences based on a threshold, noise can be effectively removed.

[0118] 2. Ensuring data consistency: In this way, the continuity of the point cloud in the physical space can be ensured. A small coordinate difference between adjacent points means that they are closely arranged in space without large jumps or breaks, which conforms to the regular physical laws.

[0119] 3. Detection of local density: This method actually detects the local density of point cloud data. By setting a difference threshold, it can identify which point cloud data is "dense" and which is "sparse" in space, thereby distinguishing which points are valid and which points may be abnormal due to noise or errors.

[0120] Thirdly, a vehicle head position interval is determined according to the minimum effective value and a preset correction value.

[0121] Optionally, the vehicle head position interval determined according to the minimum effective value and the preset correction value is:

[0122] [min_channel, min_channel+a]

[0123] Optionally, the vehicle head position interval determined according to the minimum effective value and the preset correction value is:

[0124] [min_channel-a, min_channel+a]

[0125] Wherein, min_channel is the minimum effective value, and a is the preset correction value. In order to distinguish from other correction values in the specification, a can be referred to as a third correction value. Optionally, a is set to 0.2 cm.

[0126] Fourthly, an average value of X-axis coordinate values of the second type of candidate points whose X-axis coordinate values are in the vehicle head position interval is determined as a boundary position of the vehicle head of the unmanned truck.

[0127] For a plurality of second type of candidate points, it is determined whether the X-axis coordinate value of each second type of candidate point is in the vehicle head position interval. For the second type of candidate points whose X-axis coordinate values are in the vehicle head position interval, an average value of the X-axis coordinate values of these second type of candidate points is determined, and the average value is determined as the boundary position of the vehicle head of the unmanned truck.

[0128] B4: The difference between the X-axis coordinate value of each second type of candidate point and the boundary position of the vehicle head is determined respectively, and the second type of candidate point whose difference value is less than the second threshold value is determined as the first type of effective point.

[0129] For each second-type candidate point, determine the difference between the X-axis coordinate value of the second-type candidate point and the boundary position of the vehicle head. If the difference is less than a second threshold, the second-type candidate point is determined as a first-type valid point. If the difference is greater than or equal to the second threshold, the second-type candidate point is discarded. Optionally, the second threshold is set to 5 cm.

[0130] Correspondingly, the X-axis calibration value of the packing project is determined based on the X-axis coordinate values of the first-type valid points in the M-frame point cloud data. The following scheme is adopted: the average value of the X-axis coordinate values of the first-type valid points in the M-frame point cloud data is determined as the X-axis calibration value of the packing project. That is, in the case of the work project being a packing project, the X-axis calibration value of the work project is determined based on the X-axis coordinate values of the first-type valid points in the M-frame point cloud data, which includes: determining the average value of the X-axis coordinate values of the first-type valid points in the M-frame point cloud data as the X-axis calibration value of the work project.

[0131] In the above embodiments of the present application, in the case of the work project being a packing project, the scheme for determining the first-type valid points and the scheme for determining the X-axis calibration value of the packing project based on the first-type valid points have the following technical advantages:

[0132] 1. High-precision point cloud data filtering. By obtaining the second point set generated by the vehicle head of the unmanned truck from the point cloud data and determining the point with the minimum X-axis coordinate value as the second-type candidate point for each channel of the multi-line laser radar, more representative and accurate points can be effectively filtered from a large amount of point cloud data.

[0133] 2. Adaptive difference-threshold strategy. By setting the second-type candidate point with a difference between the X-axis coordinate value and the boundary position of the vehicle head less than the second threshold as the first-type valid point, the filtering standard can be adaptively adjusted according to the changes of the actual point cloud data. For example, by dynamically adjusting the second threshold (such as 5 cm), the applicability of the algorithm can be optimized in different environments and different situations, and the robustness and adaptability of the system can be improved. This flexibility helps to handle different precision requirements in practical applications.

[0134] 3. Adapt to different laser radar types and environmental changes. In this scheme, the different characteristics of multiple channels of the multi-line laser radar are explicitly considered, so that the scheme can adapt to different types of laser radar devices. At the same time, by flexibly adjusting various thresholds, the point cloud data processing flow can be optimized in different environments (such as light, weather, etc.), and the applicable range of the algorithm can be improved.

[0135] Next, how to obtain the second-type valid points corresponding to the work project from the point cloud data will be described.

[0136] In the foregoing, in the case where the work item is the unloading item, the second type of valid point belongs to the point generated by the side of the container carried by the unmanned forklift, and in the case where the work item is the loading item, the second type of valid point belongs to the point generated by the side of the chassis of the unmanned forklift. In addition, if the multi-line laser radar is located above the middle of the lane, the multi-line laser radar can detect the two sides of the unmanned forklift / container, and if the multi-line laser radar is located on one side of the lane, the multi-line laser radar can only detect the side of the unmanned forklift / container close to the multi-line laser radar. Based on this, different ways can be used to determine the second type of valid point corresponding to the unloading item and the loading item according to different installation positions of the multi-line laser radar. The following will be described respectively.

[0137] In another embodiment of the present application, in the case where the work item is the unloading item and the multi-line laser radar is located above the middle of the lane, the second type of valid point corresponding to the unloading item is obtained from any one frame of point cloud data, and the following scheme is used:

[0138] The points generated by the two sides of the container are obtained from the point cloud data, and the points generated by the two sides of the container are determined as the second type of valid point.

[0139] Correspondingly, the Z-axis calibration value of the unloading item is determined based on the Z-axis coordinate values of the second type of valid points in M frames of point cloud data, and the following scheme is used: the average value of the Z-axis coordinate values of the second type of valid points in M frames of point cloud data is determined as the calibration value of the unloading item in the second direction. That is, in the case where the work item is the unloading item and the multi-line laser radar is located above the middle of the lane, the Z-axis calibration value of the work item is determined based on the Z-axis coordinate values of the second type of valid points in M frames of point cloud data, which includes: the average value of the Z-axis coordinate values of the second type of valid points in M frames of point cloud data is determined as the Z-axis calibration value of the work item.

[0140] Optionally, the points generated by the two sides of the container are obtained from the point cloud data, and the following scheme is used:

[0141] 1) According to the Z-axis coordinate value interval of the current work lane, the point cloud data generated by the current work lane, the unmanned forklift located in the work lane, and the container carried by the unmanned forklift is obtained from the point cloud data.

[0142] 2) The height of the front of the unmanned forklift is a known quantity, and the point cloud data generated by the container is obtained from the point cloud data obtained in the previous step based on the height of the front of the unmanned forklift.

[0143] 3) For the point cloud data generated by the container, further filter out the points generated by the front end face and the top face of the container to obtain the points generated by the two side faces of the container, and determine the points generated by the two side faces of the container as the second type of valid points.

[0144] Here, the scheme for filtering out the points generated by the front end face and the top face of the container is described in detail.

[0145] First, for the point cloud data generated by the container, determine the minimum value in the X-axis coordinate value of each point, and determine the X-axis first screening interval based on the minimum value (the upper limit value and the lower limit value of the X-axis first screening interval are: the minimum value plus a fourth correction value, and the minimum value minus the fourth correction value, for example, the fourth correction value is 10 cm); for the points generated by the container, filter out the points whose X-axis coordinate values are within the X-axis first screening interval, so as to filter out the points generated by the front end face of the container.

[0146] Second, for the point cloud data obtained in the previous step, determine the maximum value in the Y-axis coordinate value of each point, and determine the Y-axis second screening interval based on the maximum value (the upper limit value and the lower limit value of the Y-axis second screening interval are: the maximum value plus a fifth correction value, and the maximum value minus the fifth correction value, for example, the fifth correction value is 10 cm); for the point cloud data obtained in the previous step, filter out the points whose Y-axis coordinate values are within the Y-axis second screening interval, so as to filter out the points generated by the top face of the container.

[0147] In the above scheme, the points generated by the front end face of the container are filtered out first, and then the points generated by the top face of the container are filtered out.

[0148] As an alternative, the points generated by the top face of the container are filtered out first, and then the points generated by the front end face of the container are filtered out. Specifically:

[0149] First, for the point cloud data generated by the container, determine the maximum value in the Y-axis coordinate value of each point, and determine the Y-axis second screening interval based on the maximum value (the upper limit value and the lower limit value of the Y-axis second screening interval are: the maximum value plus a fifth correction value, and the maximum value minus the fifth correction value); for the points obtained in the previous step, filter out the points whose Y-axis coordinate values are within the Y-axis second screening interval, so as to filter out the points generated by the top face of the container.

[0150] Second, for the point cloud data obtained in the previous step, determine the minimum value in the X-axis coordinate value of each point, and determine the X-axis first screening interval based on the minimum value (the upper limit value and the lower limit value of the X-axis first screening interval are: the minimum value plus a fourth correction value, and the minimum value minus the fourth correction value); for the point cloud data obtained in the previous step, filter out the points whose X-axis coordinate values are within the X-axis first screening interval, so as to filter out the points generated by the front end face of the container.

[0151] In another embodiment of the present application, in the case that the work item is a container loading item and the multi-line laser radar is located above the middle of the lane, the second type of valid points corresponding to the container loading item are obtained from any frame of point cloud data, and the following scheme is adopted:

[0152] The points generated by the two sides of the chassis of the unmanned truck are obtained from the point cloud data, and the points generated by the two sides of the chassis are determined as the second type of valid points.

[0153] Correspondingly, the Z-axis calibration value of the container loading item is determined based on the Z-axis coordinate values of the second type of valid points in M frames of point cloud data, and the following scheme is adopted: the average value of the Z-axis coordinate values of the second type of valid points in M frames of point cloud data is determined as the Z-axis calibration value of the container loading item. That is, in the case that the work item is a container loading item and the multi-line laser radar is located above the middle of the lane, the Z-axis calibration value of the work item is determined based on the Z-axis coordinate values of the second type of valid points in M frames of point cloud data, including: the average value of the Z-axis coordinate values of the second type of valid points in M frames of point cloud data is determined as the Z-axis calibration value of the work item.

[0154] Optionally, the points generated by the two sides of the chassis of the unmanned truck are obtained from the point cloud data, and the following scheme is adopted:

[0155] 1) According to the Z-axis coordinate value interval of the current work lane, the point cloud data generated by the current work lane and the unmanned truck located in the work lane is obtained from the point cloud data.

[0156] 2) For the point cloud data obtained in the previous step, according to the Y-axis coordinate value of each point and the pre-constructed Y-axis first screening interval, the point cloud data is processed, only the points with Y-axis coordinate values within the Y-axis first screening interval are retained, that is, the point cloud data generated by the ground is filtered out, and the point cloud data generated by the unmanned truck is retained.

[0157] 3) For the point cloud data generated by the unmanned truck obtained in the previous step, the maximum value of the Y-axis coordinate value of each point is determined, the Y-axis third screening interval is determined based on the maximum value (the upper limit value and the lower limit value of the Y-axis third screening interval are: the maximum value plus a sixth correction value, the maximum value minus a sixth correction value, for example, the sixth correction value is 15 cm), and the points with Y-axis coordinate values within the Y-axis third screening interval are filtered out, so that the points generated by the top surface of the front of the unmanned truck are filtered out, and if the unmanned truck includes a tail, the points generated by the top surface of the tail are also filtered out.

[0158] 4) For the point cloud data obtained in the previous step, determine the minimum value of the X-axis coordinate value of each point, determine the second screening interval of the X-axis based on the minimum value (the upper limit value and the lower limit value of the second screening interval of the X-axis are: the minimum value plus a seventh correction value, the minimum value minus a seventh correction value, for example, the seventh correction value is 5 cm), and filter out the points whose X-axis coordinate values are located in the second screening interval of the X-axis, thereby filtering out the points generated by the front end face of the front of the unmanned container truck, and obtaining the points generated by the two side faces of the chassis of the unmanned container truck.

[0159] In addition, if the unmanned container truck includes a trailer, after filtering out the points whose X-axis coordinate values are located in the second screening interval of the X-axis, the following processing needs to be performed: determine the maximum value of the X-axis coordinate value of each point, determine the third screening interval of the X-axis based on the maximum value (the upper limit value and the lower limit value of the third screening interval of the X-axis are: the maximum value plus an eighth correction value, the maximum value minus an eighth correction value, for example, the eighth correction value is 5 cm), and filter out the points whose X-axis coordinate values are located in the third screening interval of the X-axis, thereby filtering out the points generated by the inner side face of the trailer of the unmanned container truck, and obtaining the points generated by the two side faces of the chassis of the unmanned container truck.

[0160] In another embodiment of the present application, in the case where the work item is a box unloading item and the multi-line laser radar is located on one side of the lane, the second type of valid points corresponding to the box unloading item are obtained from any one frame of point cloud data, and the following scheme is adopted:

[0161] The points generated by one side face of the container carried by the unmanned container truck are obtained from the point cloud data, and the points generated by one side face of the container are determined as the second type of valid points.

[0162] Correspondingly, the Z-axis calibration value of the box unloading item is determined based on the Z-axis coordinate values of the second type of valid points in M frames of point cloud data, and the following scheme is adopted: the sum of the average value of the Z-axis coordinate values of the second type of valid points in M frames of point cloud data and half of the chassis width of the unmanned container truck is determined as the Z-axis calibration value of the box unloading item. The chassis width of the unmanned container truck is a known quantity. That is, in the case where the work item is a box unloading item and the multi-line laser radar is located on one side of the lane, the Z-axis calibration value of the work item is determined based on the Z-axis coordinate values of the second type of valid points in M frames of point cloud data, which includes: the sum of the average value of the Z-axis coordinate values of the second type of valid points in M frames of point cloud data and half of the chassis width of the unmanned container truck is determined as the Z-axis calibration value of the work item.

[0163] Optionally, the points generated by one side face (specifically, the side face close to the multi-line laser radar) of the container are obtained from the point cloud data, and the following scheme is adopted:

[0164] 1) According to the Z-axis coordinate value interval of the current work lane, the point cloud data generated by the current work lane, the unmanned straddle carrier located in the work lane and the container carried by the unmanned straddle carrier are obtained from the point cloud data.

[0165] 2) The height of the front of the unmanned straddle carrier is a known quantity, and the point cloud data generated by the container is obtained from the point cloud data obtained in the previous step based on the height of the front of the unmanned straddle carrier.

[0166] 3) For the point cloud data generated by the container, the points generated by the front and top surfaces of the container are further filtered out to obtain the points generated by the side surface of the container (specifically, the side surface close to the multi-line laser radar), and the points generated by the side surface of the container are determined as the second type of valid points.

[0167] For the scheme of how to filter out the points generated by the front and top surfaces of the container, please refer to the description in the foregoing.

[0168] In another embodiment of the present application, in the case where the work project is a container loading project and the multi-line laser radar is located on one side of the lane, the second type of valid points corresponding to the container loading project are obtained from any one frame of point cloud data, and the following scheme is adopted:

[0169] The points generated by one side surface of the chassis of the unmanned straddle carrier are obtained from the point cloud data, and the points generated by one side surface of the chassis of the unmanned straddle carrier are determined as the second type of valid points.

[0170] Correspondingly, the Z-axis calibration value of the container loading project is determined based on the Z-axis coordinate values of the second type of valid points in M frames of point cloud data, and the following scheme is adopted: the sum of the average value of the Z-axis coordinate values of the second type of valid points in M frames of point cloud data and half of the width of the chassis of the unmanned straddle carrier is determined as the Z-axis calibration value of the container loading project. The width of the chassis of the unmanned straddle carrier is a known quantity. That is, in the case where the work project is a container loading project and the multi-line laser radar is located on one side of the lane, the Z-axis calibration value of the work project is determined based on the Z-axis coordinate values of the second type of valid points in M frames of point cloud data, which includes: the sum of the average value of the Z-axis coordinate values of the second type of valid points in M frames of point cloud data and half of the width of the chassis of the unmanned straddle carrier is determined as the Z-axis calibration value of the work project.

[0171] Optionally, the points generated by one side surface (specifically, the side surface close to the multi-line laser radar) of the chassis of the unmanned straddle carrier are obtained from the point cloud data, and the following scheme is adopted:

[0172] 1) According to the Z-axis coordinate value interval of the current work lane, the point cloud data generated by the current work lane and the unmanned straddle carrier located in the work lane are obtained from the point cloud data.

[0173] 2), for the point cloud data obtained in the previous step, according to the Y-axis coordinate value of each point and the first screening interval of the Y-axis constructed in advance, the point cloud data is processed, only the points whose Y-axis coordinate value is located in the first screening interval of the Y-axis are retained, that is, the point cloud data generated by the ground is filtered out, and the point cloud data generated by the unmanned container truck is retained.

[0174] 3), for the point cloud data generated by the unmanned container truck obtained in the previous step, the maximum value of the Y-axis coordinate value of each point is determined, based on the maximum value, the third screening interval of the Y-axis is determined (the upper limit value and the lower limit value of the third screening interval of the Y-axis are: the maximum value plus the sixth correction value, the maximum value minus the sixth correction value, for example, the sixth correction value is 15 cm), the points whose Y-axis coordinate value is located in the third screening interval of the Y-axis are filtered out, thereby the points generated by the top surface of the front of the unmanned container truck are filtered out, if the unmanned container truck includes the tail, the points generated by the top surface of the tail are also filtered out.

[0175] 4), for the point cloud data obtained in the previous step, the minimum value of the X-axis coordinate value of each point is determined, based on the minimum value, the second screening interval of the X-axis is determined (the upper limit value and the lower limit value of the second screening interval of the X-axis are: the minimum value plus the seventh correction value, the minimum value minus the seventh correction value, for example, the seventh correction value is 5 cm), the points whose X-axis coordinate value is located in the second screening interval of the X-axis are filtered out, thereby the points generated by the front end surface of the front of the unmanned container truck are filtered out, thereby the points generated by the side surface of the chassis of the unmanned container truck (specifically, the side surface close to the multi-line laser radar) are obtained.

[0176] In addition, if the unmanned container truck includes the tail, after the points whose X-axis coordinate value is located in the second screening interval of the X-axis are filtered out, the following processing needs to be performed: the maximum value of the X-axis coordinate value of each point is determined, based on the maximum value, the third screening interval of the X-axis is determined (the upper limit value and the lower limit value of the third screening interval of the X-axis are: the maximum value plus the eighth correction value, the maximum value minus the eighth correction value, for example, the eighth correction value is 5 cm), the points whose X-axis coordinate value is located in the third screening interval of the X-axis are filtered out, thereby the points generated by the inner side surface of the tail of the unmanned container truck are filtered out, thereby the points generated by the side surface of the chassis of the unmanned container truck (specifically, the side surface close to the multi-line laser radar) are obtained.

[0177] In another embodiment of the present application, in the case where the work item is a container loading item and the multi-line laser radar is located on one side of the lane, the second type of valid points corresponding to the container loading item are obtained from any one frame of point cloud data, the following scheme is adopted:

[0178] C1: obtain a third point set generated by the side surface of the chassis of the unmanned container truck from the point cloud data.

[0179] Since the multi-line laser radar is located on one side of the lane, the multi-line laser radar can only detect the side of the chassis of the unmanned truck close to itself. The points generated by the side of the chassis close to the multi-line laser radar can be obtained from the point cloud data, which can be referred to the scheme described in the foregoing.

[0180] C2: Determine the projection point of each point in the third point set on the X axis respectively.

[0181] It can be understood that the projection point of any one point in the third point set on the X axis has the same X axis coordinate value as the X axis coordinate value of the point in the third point set corresponding to the projection point, and the Y axis coordinate value and the Z axis coordinate value of the projection point are both 0.

[0182] C3: The radius filtering algorithm is used to filter the projection points, so as to obtain the projection points corresponding to the points satisfying the vertical surface feature, and the points in the third point set corresponding to the retained projection points are taken as the third type of candidate points.

[0183] The vertical surface refers to a plane perpendicular or approximately perpendicular to the ground in the 3D space (approximately perpendicular to the ground can be understood as: the included angle between the plane and the ground is within a first angle range, for example, 85° to 90°, including the end point value), which usually represents the side or vertical surface of an object. In this scheme, the vertical surface refers to the surface of the side of the chassis of the unmanned truck.

[0184] The non-vertical surface refers to the point cloud that does not meet the vertical surface feature. In this scheme, the non-vertical surface usually includes the front end surface of the vehicle head and the top surface of the vehicle head. Due to the structure of the front end surface and the top surface of the vehicle head, the point cloud generated by the front end surface and the top surface of the vehicle head will not be distributed along the vertical direction, so it does not meet the vertical surface feature.

[0185] The radius filtering algorithm for the projection points is a way of screening the point cloud according to the distance threshold. The radius filtering algorithm calculates the neighbor points of each point within a certain radius range, and removes those points that do not meet the requirements according to the distance. Optionally, in the radius filtering algorithm, the radius threshold is set to 3 cm, and the point cloud threshold is set to 6. Only when the number of points within 3 cm reaches 6 or more, these points are saved, and other points are removed. That is, if a point has at least 5 neighbor points within 3 cm, the point is retained, otherwise the point is removed.

[0186] Since there is a spatial relationship between the projection points and the original point cloud, the farther the projection point is from the vertical surface of the vehicle body, the more the point cloud on the vertical surface of the vehicle body will deviate. Therefore, the radius filtering algorithm will retain the projection points close to the vertical surface of the vehicle body, while remove those projection points far from the vertical surface of the vehicle body or in the vehicle head and roof area. Finally, the corresponding point cloud of the retained projection points is the point that meets the vertical surface feature.

[0187] C4: clustering the projection points of the third type of candidate points in XOZ plane, filtering out the projection points generated by the anti-collision column, and determining the third type of candidate points corresponding to the retained projection points as the second type of effective points.

[0188] Optionally, the projection points of the third type of candidate points in XOZ plane are clustered based on the Euclidean clustering algorithm, and the points with close distances are classified into one class. In this way, the projection points generated by different objects can be effectively separated. Once the clustering is completed, some projection points from the outside of the vehicle body and irrelevant to the target object (such as the projection points generated by the anti-collision column) may appear in the clustering result. The projection points generated by the anti-collision column are filtered out from the projection points of the third type of candidate points in XOZ plane, and the third type of candidate points corresponding to the retained projection points are determined as the second type of effective points.

[0189] Correspondingly, the Z-axis calibration value of the container loading project is determined based on the Z-axis coordinate values of the second type of effective points in the M-frame point cloud data, and the following scheme is adopted:

[0190] First, for each frame of point cloud data in the M-frame point cloud data, the projection points of the second type of effective points in XOZ plane are determined, and a straight line is fitted based on the projection points of the second type of effective points in XOZ plane.

[0191] Second, for each frame of point cloud data in the M-frame point cloud data, the distance of each second type of effective point in XOZ plane to the fitted straight line is calculated.

[0192]

[0193] In the above formula, x pol is the X-axis coordinate value of the projection point of the second type of effective point in XOZ plane, z pol is the Z-axis coordinate value of the projection point of the second type of effective point in XOZ plane, x ld is the X-axis vector of the fitted straight line, z ld is the Z-axis vector of the fitted straight line, and z1 is the distance of the projection point of the second type of effective point in XOZ plane to the fitted straight line.

[0194] Third, the average value of the distances of the projection points of the second type of effective points in XOZ plane in the M-frame point cloud data to the corresponding fitted straight lines and the half of the chassis width are determined as the Z-axis calibration value of the container loading project.

[0195] In the case that M is 1, a projection point of the second type of valid points in the frame of point cloud data in XOZ plane is determined, a fitting straight line is obtained by performing straight line fitting on the projection points of the second type of valid points in XOZ plane, and then an average value of distances from each projection point of the second type of valid points in the frame of point cloud data in XOZ plane to the fitting straight line is determined as the Z-axis calibration value of the packing item.

[0196] In the case that M is greater than 1, for each frame of point cloud data, the following steps are performed: a projection point of the second type of valid points in the frame of point cloud data in XOZ plane is determined, a fitting straight line is obtained by performing straight line fitting on the projection points of the second type of valid points in XOZ plane, and then distances from each projection point of the second type of valid points in the frame of point cloud data in XOZ plane to the fitting straight line are calculated. Then, an average value of distances from the projection points of the second type of valid points in M frames of point cloud data in XOZ plane to the corresponding fitting straight lines is calculated, and the average value plus half of the chassis width is determined as the Z-axis calibration value of the packing item.

[0197] Optionally, the RANSAC algorithm (random sample consensus) is used to perform straight line fitting on the projection points of the second type of valid points in the point cloud data in XOZ plane.

[0198] In the process of performing straight line fitting on the projection points of the second type of valid points in the point cloud data in XOZ plane by using the RANSAC algorithm, two projection points of the second type of valid points in XOZ plane are randomly selected to establish a straight line, distances from other projection points of the second type of valid points in XOZ plane to the straight line are calculated, points with distances less than a preset threshold value are determined as inliers, and the number of inliers is recorded; another two projection points of the second type of valid points in XOZ plane are selected to establish a straight line, distances from other projection points of the second type of valid points in XOZ plane to the straight line are calculated, points with distances less than a preset threshold value are determined as inliers, and the number of inliers is recorded; the above steps are repeatedly performed for multiple times, the straight line with the largest number of inliers is taken as the best fitting straight line, and the fitting process is completed.

[0199] The applicant considers that if the Z-axis calibration is only based on one side of the chassis, there may be deviation due to the unevenness of the side of the chassis of the unmanned truck, and proposes the technical solution described above. In the embodiments disclosed above in the present application, for the third point set generated by one side of the unmanned truck, a radius filtering algorithm is used to filter the projection points of each point in the X-axis, so as to obtain points satisfying the vertical plane feature, and these points are used as the third type of candidate points. Then, the projection points of the third type of candidate points in the XOZ plane are clustered, so as to filter out the third type of candidate points generated by the anti-collision column, and the remaining third type of candidate points are determined as the second type of effective points. Then, the projection points of the second type of effective points in the XOZ plane are fitted with a straight line, and the Z-axis calibration value of the loading project is determined based on the distance between the projection points of the second type of effective points in the XOZ plane and the fitted straight line and the width of the chassis, which can eliminate the deviation caused by the unevenness of the side of the chassis, and make the calibration result more accurate.

[0200] It should be noted that in the foregoing, the negative direction of the X-axis is taken as the forward direction of the unmanned truck in the lane, the positive direction of the Y-axis is taken as the vertical upward direction, and the positive direction of the Z-axis is taken as the left side of the lane. It can be understood that it is also possible to define the positive direction of one or more of the X-axis, Y-axis and Z-axis as another direction, and the principle is the same.

[0201] For example, the positive direction of the X-axis is defined as the forward direction of the unmanned truck in the lane.

[0202] Correspondingly, in the case of the work project being the unloading project, the first type of effective points corresponding to the work project are obtained from a frame of point cloud data, including:

[0203] The first point set generated by the container is obtained from the point cloud data;

[0204] The first maximum value in the X-axis coordinate values of the points in the first point set is determined;

[0205] For each channel of the multiple channels of the multi-line laser radar, among the points in the first point set belonging to the channel, the point with the maximum X-axis coordinate value is determined as the first type of candidate point;

[0206] The difference between the X-axis coordinate value of each first type of candidate point and the first maximum value is determined, and the first type of candidate point with a difference less than a first threshold value is determined as the first type of effective point.

[0207] Correspondingly, in the case of the work project being the loading project, the first type of effective points corresponding to the work project are obtained from a frame of point cloud data, including:

[0208] The second point set is obtained from the point cloud data, and the second point set contains points generated by the front of the unmanned truck;

[0209] For each channel of the multiple channels of the multi-line laser radar, among the multiple points belonging to the channel in the second point set, a point with the maximum X-axis coordinate value is determined as a second-type candidate point;

[0210] The boundary position of the vehicle head of the unmanned container truck is determined based on the second-type candidate points;

[0211] A difference value between the X-axis coordinate value of each second-type candidate point and the boundary position is determined, and a second-type candidate point with a difference value less than a second threshold value is determined as a first-type valid point.

[0212] Correspondingly, the boundary position of the vehicle head of the unmanned container truck is determined based on the second-type candidate points, including:

[0213] The second-type candidate points are sorted according to the X-axis coordinate values;

[0214] A maximum effective value is determined based on the X-axis coordinate values of the sorted multiple second-type candidate points, wherein the maximum effective value is a maximum value among all X-axis coordinate values contained in all second-type candidate point pairs whose adjacent positions after sorting and X-axis coordinate value difference are less than a third threshold value;

[0215] A vehicle head position interval is determined according to the maximum effective value and a preset correction value;

[0216] An average value of the X-axis coordinate values of the second-type candidate points whose X-axis coordinate values are in the vehicle head position interval is determined as the boundary position of the vehicle head of the unmanned container truck.

[0217] It can be understood that if the positive direction of the X-axis is defined as the forward direction of the unmanned container truck in the lane, the configuration mode of the upper limit value and the lower limit value of each X-axis screening interval used in the scheme for obtaining the first-type valid point and the second-type valid point needs to be adaptively adjusted, but the principle is the same; if the positive direction of the Y-axis is defined as vertically downward, the configuration mode of the upper limit value and the lower limit value of each Y-axis screening interval used in the scheme for obtaining the first-type valid point and the second-type valid point needs to be adaptively adjusted, but the principle is the same; if the positive direction of the Z-axis is defined as the right side of the lane, the configuration mode of the upper limit value and the lower limit value of each Z-axis screening interval used in the scheme for obtaining the first-type valid point and the second-type valid point remains unchanged.

[0218] Here is an example.

[0219] In the case where the positive direction of the X-axis is defined as the forward direction of the unmanned container truck in the lane, and the positive direction of the Y-axis is defined as vertically downward, the scheme for filtering out the points generated by the front end face and the top face of the container includes:

[0220] First, for the point cloud data generated by the container, the maximum value in the X-axis coordinate value of each point is determined, and the first screening interval of the X-axis is determined based on the maximum value (the upper limit value and the lower limit value of the first screening interval of the X-axis are: the maximum value plus the fourth correction value, and the maximum value minus the fourth correction value); for the points generated by the container, the points with the X-axis coordinate value in the first screening interval of the X-axis are filtered out, so that the points generated by the front end face of the container are filtered out.

[0221] Second, for the point cloud data obtained in the previous step, the minimum value in the Y-axis coordinate value of each point is determined, and the second screening interval of the Y-axis is determined based on the minimum value (the upper limit value and the lower limit value of the second screening interval of the Y-axis are: the minimum value plus the fifth correction value, and the maximum value minus the fifth correction value); for the point cloud data obtained in the previous step, the points with the Y-axis coordinate value in the second screening interval of the Y-axis are filtered out, so that the points generated by the top surface of the container are filtered out.

[0222] In the case where the positive direction of the X-axis is defined as the advancing direction of the unmanned container truck in the lane, and the positive direction of the Y-axis is defined as the vertical downward direction, the scheme for obtaining the points generated by the two side surfaces of the chassis of the unmanned container truck from the point cloud data includes:

[0223] 1) According to the Z-axis coordinate value interval of the current working lane, the point cloud data generated by the current working lane and the unmanned container truck located in the working lane is obtained from the point cloud data.

[0224] 2) For the point cloud data obtained in the previous step, the Y-axis coordinate value of each point and the first screening interval of the Y-axis constructed in advance are used to process the point cloud data, and only the points with the Y-axis coordinate value in the first screening interval of the Y-axis are retained, that is, the point cloud data generated by the ground in the point cloud data is filtered out, and the point cloud data generated by the unmanned container truck is retained.

[0225] 3) For the point cloud data generated by the unmanned container truck obtained in the previous step, the minimum value of the Y-axis coordinate value of each point is determined, and the third screening interval of the Y-axis is determined based on the minimum value (the upper limit value and the lower limit value of the third screening interval of the Y-axis are: the minimum value plus the sixth correction value, and the maximum value minus the sixth correction value), and the points with the Y-axis coordinate value in the third screening interval of the Y-axis are filtered out, so that the points generated by the top surface of the vehicle head of the unmanned container truck are filtered out, and if the unmanned container truck includes a vehicle tail, the points generated by the top surface of the vehicle tail are also filtered out.

[0226] 4) For the point cloud data obtained in the previous step, determine the maximum value of the X-axis coordinate value of each point, determine the X-axis second screening interval based on the maximum value (the upper limit value and the lower limit value of the X-axis second screening interval are: the maximum value plus the seventh correction value, the maximum value minus the seventh correction value), and filter out the points whose X-axis coordinate values are located in the X-axis second screening interval, thereby filtering out the points generated by the front end face of the front of the unmanned container truck, and obtaining the points generated by the two side faces of the chassis of the unmanned container truck.

[0227] In addition, if the unmanned container truck includes a trailer, after filtering out the points whose X-axis coordinate values are located in the X-axis second screening interval, the following processing needs to be performed: determine the minimum value of the X-axis coordinate value of each point, determine the X-axis third screening interval based on the minimum value (the upper limit value and the lower limit value of the X-axis third screening interval are: the minimum value plus the eighth correction value, the minimum value minus the eighth correction value), and filter out the points whose X-axis coordinate values are located in the X-axis third screening interval, thereby filtering out the points generated by the inner side face of the trailer of the unmanned container truck, and obtaining the points generated by the two side faces of the chassis of the unmanned container truck.

[0228] In the above-mentioned embodiments of the present application, the point cloud data obtained by the electronic device can be in CSV format. That is, the multi-line laser radar performs scanning to generate point cloud data in PCD format, saves the point cloud data in PCD format as a CSV file, and the electronic device obtains point cloud data in CSV format during the process of performing the calibration method disclosed in the present solution. Specifically, one frame of point cloud data is saved as one CSV file. The electronic device obtains M CSV files and calibrates the job location of the job item based on the M CSV files.

[0229] A row in the CSV file is the data of one point, and the data of one point includes the three-dimensional coordinates of the point and the laser reflection intensity. Taking a 64-line laser radar as an example, assuming that the horizontal resolution of the 64-line laser radar is 0.2, and the driving mechanism drives the laser transceiver module to rotate 360° in one scanning period, then the number of points collected by the 64-line laser radar in one scanning period is 115200, and these points constitute one frame of point cloud data, and accordingly, one CSV file includes 115200 rows.

[0230] In the present solution, the electronic device obtains point cloud data in CSV format, which has the following advantages:

[0231] First, the point cloud data in CSV format is easy to parse and edit.

[0232] CSV files are highly readable: CSV files are in plain text format and can be viewed and edited using various text editors. CSV files are highly compatible: Many data processing and analysis tools natively support CSV files. CSV files are easy to process: The structure of CSV files is simple and can be easily integrated with programming languages.

[0233] Second, the point cloud data in CSV format facilitates channel division.

[0234] The data of multiple points generated by one channel of a multi-line laser radar are adjacent in the CSV file. Therefore, based on the point cloud data in CSV format, channel division can be simply performed.

[0235] For example, the horizontal resolution of a 64-line laser radar is 0.2, and in one scanning period, the driving mechanism drives the laser transceiver module to rotate 360°. Therefore, in one scanning period, 1800 points are collected by each channel of the 64-line laser radar, and in the CSV file, the data of the points collected by the first channel are in the first row to the 1800th row, the data of the points collected by the second channel are in the 1801st row to the 3600th row, and the data of the points collected by other channels are in the CSV file in a similar manner.

[0236] Third, the electronic device can batch process multiple CSV files.

[0237] After the calibration of the work item based on the technical solution disclosed in the present application, the scheme for controlling the unmanned truck during its travel based on the calibration result is as follows:

[0238] Based on the point cloud data generated by the multi-line laser radar, the actual coordinate values of the unmanned truck on the X-axis and the Z-axis are determined.

[0239] The X-axis calibration value and the Z-axis calibration value corresponding to the current work item are obtained.

[0240] The position of the unmanned truck on the X-axis is controlled based on the X-axis calibration value and the actual coordinate value of the unmanned truck on the X-axis, and the position of the unmanned truck on the Z-axis is controlled based on the Z-axis calibration value and the actual coordinate value of the unmanned truck on the Z-axis.

[0241] It can be understood that the scheme for determining the actual coordinate value of the unmanned truck on the X-axis is similar to the scheme for determining the X-axis calibration value during calibration, and the scheme for determining the actual coordinate value of the unmanned truck on the Z-axis is similar to the scheme for determining the Z-axis calibration value during calibration.

[0242] Briefly described here: obtain a frame of point cloud data collected by a multi-line laser radar; obtain first type of valid points and second type of valid points corresponding to a current work item from the frame of point cloud data; determine actual coordinate values of the unmanned straddle carrier in the X axis based on X axis coordinate values of the first type of valid points, and determine actual coordinate values of the unmanned straddle carrier in the Z axis based on Z axis coordinate values of the second type of valid points.

[0243] In the above control scheme, the scheme of obtaining the first type of valid points and the second type of valid points corresponding to the current work item from the point cloud data is similar to the scheme of obtaining the first type of valid points and the second type of valid points corresponding to the work item in the calibration process. The scheme of determining the actual coordinate values of the unmanned straddle carrier in the X axis based on the X axis coordinate values of the first type of valid points is similar to the scheme of determining the X axis calibration value of the work item based on the X axis coordinate values of the first type of valid points in the calibration process when M is 1. The scheme of determining the actual coordinate values of the unmanned straddle carrier in the Z axis based on the Z axis coordinate values of the second type of valid points is similar to the scheme of determining the Z axis calibration value of the work item based on the Z axis coordinate values of the second type of valid points in the calibration process when M is 1.

[0244] The application also discloses a calibration device of an unmanned straddle carrier positioning system, which has the structure as shown in the figure. Figure 3 The calibration device comprises an indication information acquisition module 301, a point cloud data acquisition module 302, a first type of valid point acquisition module 303, a first calibration module 304, a second type of valid point acquisition module 305, and a second calibration module 306.

[0245] The indication information acquisition module 301 is used for obtaining indication information, and the indication information comprises a work item and a work lane.

[0246] The point cloud data acquisition module 302 is used for obtaining M frames of point cloud data, which are collected by the multi-line laser radar when the unmanned straddle carrier is parked at an expected position of the work item in the work lane, and M is greater than or equal to 1.

[0247] The first type of valid point acquisition module 303 is used for obtaining the first type of valid points corresponding to the work item from each frame of point cloud data, wherein, in the case that the work item is an unloading item, the first type of valid points belong to points generated by the front end face of the container carried by the unmanned straddle carrier, and in the case that the work item is a loading item, the first type of valid points belong to points generated by the front end of the vehicle head of the unmanned straddle carrier.

[0248] The first calibration module 304 is used for determining the X axis calibration value of the work item based on the X axis coordinate values of the first type of valid points in the M frames of point cloud data.

[0249] The second type of valid point acquisition module 305 is used to obtain the second type of valid points corresponding to the operation project from each frame of point cloud data. In the case of the operation project being a container unloading project, the second type of valid points are points generated on the side of the container carried by the unmanned truck. In the case of the operation project being a container loading project, the second type of valid points are points generated on the side of the chassis of the unmanned truck.

[0250] The second calibration module 306 is used to determine the Z-axis calibration value of the work item based on the Z-axis coordinate values ​​of the second type of valid points in the M-frame point cloud data.

[0251] Wherein, the X-axis represents the length of the lane, and the Z-axis represents the width of the lane.

[0252] The specific implementation of each module in the calibration device can be found in the description of the corresponding steps of the calibration method above, and will not be repeated here.

[0253] This application also discloses an electronic device.

[0254] See Figure 4 , Figure 4 The hardware structure of an electronic device is shown, which includes a processor 401, a memory 402, a communication interface 403, and a communication bus 404.

[0255] In this embodiment, the number of processor 401, memory 402, communication interface 403, and communication bus 404 is at least one, and the processor 401, memory 402, and communication interface 403 communicate with each other through communication bus 404. Communication bus 404 can be an ISA (Industry Standard Architecture) bus, a PCI (Peripheral Component Interconnect) bus, or an EISA (Extended Industry Standard Architecture) bus, etc. The bus can be divided into address bus, data bus, control bus, etc.

[0256] It should be noted that those skilled in the art will understand that Figure 4 The structure of the electronic device shown does not constitute a limitation on the electronic device; the electronic device may include, but is not limited to, other electronic devices. Figure 4 This may indicate more or fewer components, or combinations of certain components, or different component arrangements.

[0257] The following is combined Figure 4 A detailed introduction to each component of the electronic device is provided.

[0258] The processor 401 is the control center of the electronic device, connects each part of the whole electronic device by various interfaces and lines, executes various functions of the electronic device and processes data by running or executing software programs and / or modules stored in the memory 402 and calling data stored in the memory 402, thereby monitoring the whole electronic device.

[0259] The processor 401 can be a central processing unit (CPU), or an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement embodiments of the present application, etc.

[0260] The memory 402 can include memory, such as random-access memory (RAM) and read-only memory (ROM), and can also include a mass storage device, such as at least one disk memory, etc.

[0261] The memory 402 stores a computer program, and the processor 401 is configured to execute the computer program stored in the memory 402, so that the electronic device implements any one of the calibration methods disclosed above.

[0262] The present application also provides a readable storage medium, the readable storage medium carries one or more computer programs, when the one or more computer programs are executed by the electronic device, so that the electronic device implements any one of the calibration methods disclosed above.

[0263] It should be noted that the technical features of each embodiment in the specification can be replaced or combined with each other, and each embodiment focuses on the difference from other embodiments, and the same or similar parts of each embodiment can be referred to each other. The steps in the method of each embodiment can be adjusted, combined and deleted according to actual needs. The modules in the device and equipment of each embodiment can be combined, divided and deleted according to actual needs. For the calibration device of the unmanned container truck positioning system, the electronic device and the readable storage medium disclosed in the embodiments, since they correspond to the calibration method of the unmanned container truck positioning system disclosed in the embodiments, the description is relatively simple, and the related parts are referred to the method part.

[0264] In this document, relational terms such as first and second and the like can be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Moreover, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but can include other elements not expressly listed or inherent to such process, method, article, or apparatus. An element proceeded by "comprises... a" does not, without more constraints, exclude the existence of additional identical elements in the process, method, article, or apparatus that comprises the element.

[0265] The above description of disclosed embodiments provides enabling concepts for practicing or using the application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein can be applied to other embodiments without departing from the spirit or scope of the application. Thus, the present application is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for calibrating an unmanned container truck positioning system, characterized in that, The method comprises the following steps: obtaining indication information, the indication information comprising a work item and a work lane; obtaining M frames of point cloud data collected by a multi-line laser radar when an unmanned truck stops at a desired position of the work item in the work lane, M being greater than or equal to 1; obtaining, from each frame of point cloud data, a first type of valid point corresponding to the work item, wherein, in the case that the work item is an unloading item, the first type of valid point belongs to a point generated by a front end face of a container carried by the unmanned truck, and in the case that the work item is a loading item, the first type of valid point belongs to a point generated by a front end of a truck head of the unmanned truck; determining an X-axis calibration value of the work item based on X-axis coordinate values of the first type of valid points in the M frames of point cloud data; obtaining, from each frame of point cloud data, a second type of valid point corresponding to the work item, wherein, in the case that the work item is an unloading item, the second type of valid point belongs to a point generated by a side face of the container carried by the unmanned truck, and in the case that the work item is a loading item, the second type of valid point belongs to a point generated by a side face of a chassis of the unmanned truck; determining a Z-axis calibration value of the work item based on Z-axis coordinate values of the second type of valid points in the M frames of point cloud data; wherein the X-axis is a length direction of the lane, and the Z-axis is a width direction of the lane.

2. The method of claim 1, wherein, A negative direction of the X-axis is a forward direction of the unmanned truck in the lane; in the case that the work item is an unloading item, obtaining, from a frame of point cloud data, a first type of valid point corresponding to the work item comprises: obtaining, from the point cloud data, a first point set generated by the container; determining a first minimum value in X-axis coordinate values of points in the first point set; for each channel of a plurality of channels of the multi-line laser radar, determining, in the points in the first point set belonging to the channel, a point with a minimum X-axis coordinate value as a first type of candidate point; respectively determining a difference between an X-axis coordinate value of each first type of candidate point and the first minimum value, and determining a first type of candidate point with a difference less than a first threshold value as a first type of valid point; in the case that the work item is an unloading item, the determining of the X-axis calibration value of the work item based on the X-axis coordinate values of the first type of valid points in the M frames of point cloud data comprises: determining an average value of the X-axis coordinate values of the first type of valid points in the M frames of point cloud data as the X-axis calibration value of the work item.

3. The method of claim 1, wherein, A negative direction of the X-axis is a forward direction of the unmanned truck in the lane; in the case that the work item is a loading item, obtaining, from a frame of point cloud data, a first type of valid point corresponding to the work item comprises: obtaining, from the point cloud data, a second point set containing points generated by a truck head of the unmanned truck; for each channel of a plurality of channels of the multi-line laser radar, determining, in a plurality of points in the second point set belonging to the channel, a point with a minimum X-axis coordinate value as a second type of candidate point; determining a boundary position of the truck head of the unmanned truck based on the second type of candidate point; Determine the difference between the X-axis coordinate value of each second-type candidate point and the boundary position, and determine the second-type candidate point with a difference less than a second threshold as a first-type valid point; In a case where the work item is a loading item, the X-axis calibration value of the work item is determined based on the X-axis coordinate values of the first-type valid points in the M frames of point cloud data, including: determining the average value of the X-axis coordinate values of the first-type valid points in the M frames of point cloud data as the X-axis calibration value of the work item.

4. The method of claim 3, wherein, Determine the boundary position of the front of the unmanned truck based on the second-type candidate points, including: Sort the second-type candidate points according to the X-axis coordinate values; Determine a minimum effective value based on the X-axis coordinate values of the sorted second-type candidate points, wherein the minimum effective value is the minimum value of all X-axis coordinate values contained in all second-type candidate point pairs that are adjacent in position after sorting and have a difference in X-axis coordinate values less than a third threshold; Determine a front position interval based on the minimum effective value and a preset correction value; Determine the average value of the X-axis coordinate values of the second-type candidate points in the front position interval as the boundary position of the front of the unmanned truck.

5. The method of claim 1, wherein, In a case where the work item is an unloading item and the multi-line laser radar is located above the middle of the lane, the second-type valid points corresponding to the work item are obtained from a frame of point cloud data, including: obtaining points generated by two side surfaces of the container from the point cloud data, and determining the points generated by the two side surfaces of the container as the second-type valid points; In a case where the work item is a loading item and the multi-line laser radar is located above the middle of the lane, the second-type valid points corresponding to the work item are obtained from a frame of point cloud data, including: obtaining points generated by two side surfaces of the chassis of the unmanned truck from the point cloud data, and determining the points generated by the two side surfaces of the chassis as the second-type valid points; In a case where the multi-line laser radar is located above the middle of the lane, the Z-axis calibration value of the work item is determined based on the Z-axis coordinate values of the second-type valid points in the M frames of point cloud data, including: determining the average value of the Z-axis coordinate values of the second-type valid points in the M frames of point cloud data as the Z-axis calibration value of the work item.

6. The method of claim 1, wherein, In a case where the work item is an unloading item and the multi-line laser radar is located on one side of the lane, the second-type valid points corresponding to the work item are obtained from a frame of point cloud data, including: obtaining points generated by one side surface of the container from the point cloud data, and determining the points generated by the one side surface of the container as the second-type valid points; In the case that the work item is a loading item and the multi-line laser radar is located at one side of the lane, the second type of valid points corresponding to the work item are obtained from a frame of point cloud data, including: obtaining points generated by one side of the chassis of the unmanned truck from the point cloud data, and determining the points generated by one side of the chassis of the unmanned truck as the second type of valid points. In the case that the multi-line laser radar is located at one side of the lane, the Z-axis calibration value of the work item is determined based on the Z-axis coordinate values of the second type of valid points in the M frames of point cloud data, including: determining the sum of the average value of the Z-axis coordinate values of the second type of valid points in the M frames of point cloud data and one-half of the width of the chassis of the unmanned truck as the Z-axis calibration value of the work item.

7. The method of claim 1, wherein, In the case that the work item is a loading item and the multi-line laser radar is located at one side of the lane, the second type of valid points corresponding to the work item are obtained from a frame of point cloud data, including: obtaining a third point set generated by one side of the unmanned truck from the point cloud data; determining the projection points of each point in the third point set on the X-axis respectively; filtering the projection points by using a radius filtering algorithm so as to obtain the projection points corresponding to the points satisfying the vertical plane feature, and determining the points in the third point set corresponding to the retained projection points as the third type of candidate points; clustering the projection points of the third type of candidate points on the XOZ plane, filtering out the projection points generated by the anti-collision columns, and determining the third type of candidate points corresponding to the retained projection points as the second type of valid points; In the case that the work item is a loading item and the multi-line laser radar is located at one side of the lane, the Z-axis calibration value of the work item is determined based on the Z-axis coordinate values of the second type of valid points in the M frames of point cloud data, including: for each frame of point cloud data, determining the projection points of the second type of valid points in the point cloud data on the XOZ plane, performing linear fitting on the projection points of the second type of valid points on the XOZ plane to obtain a fitting straight line, and determining the sum of the average value of the distances from the projection points of the second type of valid points in the M frames of point cloud data on the XOZ plane to the corresponding fitting straight line and one-half of the width of the chassis as the Z-axis calibration value of the work item.

8. A calibration device for an unmanned container truck positioning system, characterized in that including: an indication information acquisition module configured to obtain indication information, the indication information including a work item and a work lane; a point cloud data acquisition module configured to obtain M frames of point cloud data, the M frames of point cloud data being collected by a multi-line laser radar when an unmanned truck is parked at an expected position of the work item in the work lane, M being greater than or equal to 1; a first type of valid point acquisition module configured to obtain a first type of valid point corresponding to the work item from each frame of point cloud data, wherein in the case that the work item is an unloading item, the first type of valid point belongs to points generated by a front end of a container carried by the unmanned truck, and in the case that the work item is a loading item, the first type of valid point belongs to points generated by a front end of a truck head of the unmanned truck; The first calibration module is configured to determine an X-axis calibration value of the work item based on X-axis coordinate values of the first type of valid points in the M frames of point cloud data. The second type of valid point acquisition module is configured to obtain, from each frame of point cloud data, a second type of valid point corresponding to the work item, wherein, in a case where the work item is an unloading box item, the second type of valid point belongs to points generated on a side of a container carried by the unmanned truck, and in a case where the work item is a loading box item, the second type of valid point belongs to points generated on a side of a chassis of the unmanned truck. The second calibration module is configured to determine a Z-axis calibration value of the work item based on Z-axis coordinate values of the second type of valid points in the M frames of point cloud data. The X-axis is a length direction of a lane, and the Z-axis is a width direction of the lane.

9. An electronic device, comprising: The electronic device includes at least one processor and a memory connected to the processor, wherein: The memory is configured to store a computer program. The processor is configured to execute the computer program to enable the electronic device to implement the calibration method according to any one of claims 1 to 7.

10. A readable storage medium, characterized by, The readable storage medium carries one or more computer programs, and when the one or more computer programs are executed by an electronic device, the electronic device implements the calibration method according to any one of claims 1 to 7.

Images