WO2021179988A1 - 基于三维激光的集卡防砸检测方法、装置和计算机设备 - Google Patents
基于三维激光的集卡防砸检测方法、装置和计算机设备 Download PDFInfo
- Publication number
- WO2021179988A1 WO2021179988A1 PCT/CN2021/079102 CN2021079102W WO2021179988A1 WO 2021179988 A1 WO2021179988 A1 WO 2021179988A1 CN 2021079102 W CN2021079102 W CN 2021079102W WO 2021179988 A1 WO2021179988 A1 WO 2021179988A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- point cloud
- container
- lidar
- dimensional
- spreader
- Prior art date
Links
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S17/00—Systems using the reflection or reradiation of electromagnetic waves other than radio waves, e.g. lidar systems
- G01S17/88—Lidar systems specially adapted for specific applications
- G01S17/93—Lidar systems specially adapted for specific applications for anti-collision purposes
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B65—CONVEYING; PACKING; STORING; HANDLING THIN OR FILAMENTARY MATERIAL
- B65G—TRANSPORT OR STORAGE DEVICES, e.g. CONVEYORS FOR LOADING OR TIPPING, SHOP CONVEYOR SYSTEMS OR PNEUMATIC TUBE CONVEYORS
- B65G67/00—Loading or unloading vehicles
- B65G67/02—Loading or unloading land vehicles
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B66—HOISTING; LIFTING; HAULING
- B66C—CRANES; LOAD-ENGAGING ELEMENTS OR DEVICES FOR CRANES, CAPSTANS, WINCHES, OR TACKLES
- B66C13/00—Other constructional features or details
- B66C13/18—Control systems or devices
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B66—HOISTING; LIFTING; HAULING
- B66C—CRANES; LOAD-ENGAGING ELEMENTS OR DEVICES FOR CRANES, CAPSTANS, WINCHES, OR TACKLES
- B66C15/00—Safety gear
- B66C15/06—Arrangements or use of warning devices
Definitions
- This application relates to the field of laser radar technology, and in particular to a three-dimensional laser-based truck anti-smashing detection method, device and computer equipment.
- the truck loading operation refers to the use of the spreader of a container crane to clamp the container in the container stack, hoist it, and control its drop to be loaded and cut to the tray of the truck.
- a two-dimensional laser scanner is introduced for the truck packing function to perform detection.
- the two-dimensional laser scanner scans the container and the truck along the center line parallel to the truck lane, and calculates the deviation field of the truck in real time.
- the distance value of the lifting point of the bridge is notified to the driver through the LED screen to adjust the position of the truck, so as to achieve the purpose of anti-smashing.
- this method can only correct and adjust the truck in the front and rear directions, and the correction direction is affected by the installation position of the lidar, resulting in low accuracy of anti-smashing detection.
- a three-dimensional laser-based anti-smashing detection method for trucks includes:
- An anti-smashing detection device for trucks based on a three-dimensional laser comprising:
- the container acquisition module is used to acquire the size parameters of the container currently clamped by the spreader of the container crane in the truck loading operation;
- a point cloud acquisition module which acquires a three-dimensional point cloud of container operations collected by lidar when the spreader clamps the container and falls;
- An attitude parameter acquisition module for acquiring the attitude parameters of the lidar
- a position acquisition module for acquiring the relative translation amount of the spreader and the reference lidar
- a drop area determination module configured to determine the drop area range of the container in the comprehensive point cloud according to the relative translation amount and the size parameter of the container;
- the detection module is used to send out an anti-smashing alarm when an obstacle is detected within the falling area of the container.
- a computer device includes a memory and a processor, the memory stores a computer program, and the processor implements the steps of the methods of the foregoing embodiments when the computer program is executed.
- a computer-readable storage medium has a computer program stored thereon, and when the computer program is executed by a processor, it realizes the steps of each of the above-mentioned implementation methods below.
- the above-mentioned three-dimensional laser-based truck anti-smashing detection method, device, computer equipment and storage medium use lidar to collect three-dimensional data of container operations when the spreader clamps the container and falls.
- the data has high accuracy and is based on high-precision three-dimensional point cloud data.
- the three-dimensional point cloud is converted according to the attitude parameters to obtain a comprehensive point cloud, so that the monitoring range is not affected by the installation position of the lidar, and then according to the container size parameter clamped by the fixture, the translation amount of the spreader and the reference lidar is
- the full point cloud determines the range of the container's falling area, and when an obstacle is detected within the range of the container's falling area, an anti-smashing alarm is issued.
- the accuracy of the data source of this method is high, and the detection method is not affected by the installation position of the lidar, which greatly improves the accuracy of the anti-smashing detection.
- FIG. 1 is an application environment diagram of a three-dimensional laser-based pickup anti-smashing detection method in an embodiment
- FIG. 2 is a schematic diagram of a scene where there are no obstacles in the drop area of the container in the container packing operation in an embodiment
- FIG. 3 is a schematic diagram of a scene where there are obstacles in the drop area of the container in the container packing operation in an embodiment
- FIG. 4 is a schematic flow chart of a three-dimensional laser-based anti-smashing detection method for trucks in an embodiment
- Figure 5 is a schematic diagram of the relative positional relationship between the spreader and the reference lidar in an embodiment
- FIG. 6 is a schematic diagram of the relative position relationship between the spreader and the reference lidar in another embodiment
- Figure 7 is a schematic diagram of the relative positional relationship between the spreader and the reference lidar when the spreader lifts the container in an embodiment
- Figure 8 is a schematic diagram of a container drop area in an embodiment
- FIG. 9 is an application environment diagram of a three-dimensional laser-based pickup anti-smashing detection method in another embodiment.
- FIG. 10 is a schematic flowchart of a three-dimensional laser-based anti-smashing detection method for trucks in another embodiment
- FIG. 11 is a schematic diagram of setting the coordinate system of the detection system in an embodiment
- FIG. 12 is a schematic flowchart of the step of obtaining the first attitude angle of the reference lidar in an embodiment
- Figure 13 is a schematic flow chart of the steps of issuing an anti-smashing alarm when an obstacle is detected within the falling area of the container in an embodiment
- FIG. 14 is a schematic diagram of the relationship between a side view of a truck and its warning area and height threshold in an embodiment
- 15 is a schematic flowchart of the steps of issuing an anti-smashing alarm when an obstacle is detected within the falling area of the container in another embodiment
- FIG. 16 is a schematic flowchart of the steps of projecting a comprehensive point cloud into a two-dimensional image in an embodiment
- Figure 17 is a schematic diagram of a two-dimensional image of no obstacles under the container in an embodiment
- Figure 18 is a schematic diagram of a two-dimensional image of an obstacle under the container in an embodiment
- 19 is a flow diagram of the steps of performing image detection on the pixels in the location range of the container falling area in the two-dimensional image in an embodiment, and issuing an anti-smashing alarm if an obstacle is detected;
- FIG. 20 is a structural block diagram of an anti-smashing detection device for trucks based on a three-dimensional laser in an embodiment
- Fig. 21 is a diagram of the internal structure of a computer device in an embodiment.
- the three-dimensional laser-based anti-smashing detection method for collecting trucks provided in this application can be applied to the application environment as shown in FIG. 1.
- the lidar 101 is installed on the same side below the spreader 103 of the container crane 102 at a certain angle, and collects the three-dimensional point cloud of the container operation when the spreader clamps the container and falls.
- the installation position of the lidar is set according to the height of the truck.
- the main control device 105 is in communication connection with the lidar 101.
- the main control device is also connected to the control device 106 of the container crane 102. Both the main control device 105 and the control device 106 can be installed in the control room of the container crane.
- the control device 106 controls the spreader 103 of the container crane 102 to clamp the container 107 in the container stack, it sends the size parameter of the currently clamped container to the main control device 105.
- the control device 106 sends a signal to the main control device 105 that the spreader starts to fall.
- the main control device 105 sends a collection signal to the lidar 101 according to the signal, and the lidar 101 collects the three-dimensional point cloud of the container operation.
- the laser radar detects an obstacle with a certain height (such as the truck head, other containers on the truck tray, etc.) in the container falling area directly below the container, and outputs an anti-smashing alarm signal to Control equipment for container cranes.
- a three-dimensional laser-based anti-smashing detection method for trucks is provided.
- the method is applied to the main control device in Figure 1 as an example for description, including the following steps:
- the truck loading operation refers to the use of the spreader of a container crane to clamp the container in the container stack, hoist it, and control its drop to be loaded and cut to the tray of the truck.
- the control device of the container crane sends the container size parameter clamped by the spreader to the main control device.
- the size parameters of the container include the length, width and height of the container.
- S204 Acquire a three-dimensional point cloud of the container operation collected by the lidar when the container is clamped by the spreader and falls.
- Jida radar collects a three-dimensional laser point cloud of a container operation site.
- the control device sends a signal to the main control device that the spreader starts to fall.
- the main control device sends a collection signal to the lidar according to the signal, and the lidar collects the three-dimensional point cloud of the container operation.
- the attitude parameters include attitude angle, which refers to the installation angle of the lidar relative to the reference object, including but not limited to roll angle, pitch angle, and yaw angle.
- the attitude angle of the lidar can be determined according to the three-dimensional point cloud of the container operation.
- the attitude angle of the lidar since the position of the lidar is basically fixed after the installation of the truck anti-smashing detection device, the attitude angle of the lidar only needs to be calculated once, and the first attitude angle can be used in the subsequent point cloud calibration. Each test can be calibrated in real time, so that the calibrated point cloud will be more accurate.
- the attitude angle includes a roll angle, a pitch angle, and a yaw angle.
- the step of obtaining the attitude angle of the lidar includes: obtaining the three-dimensional calibration point cloud of the container operation collected by the lidar in the calibration state; determining the ground point cloud from the three-dimensional calibration point cloud according to the installation height of the lidar; calculating the ground point The plane normal vector of the cloud; calculate the roll angle and pitch angle of the lidar according to the plane normal vector of the ground point cloud; work from the container according to the installation height of the lidar, the height of the truck tray, the height of the container, and the distance from the lidar Determine the point cloud on the side of the container in the three-dimensional point cloud; calculate the plane normal vector of the point cloud on the side of the container; calculate the yaw angle of the lidar according to the plane normal vector of the point cloud on the side of the container.
- the attitude angle includes roll angle, pitch angle and yaw angle.
- the roll angle and pitch angle are obtained according to the plane normal vector of the ground point cloud in the 3D point cloud
- the yaw angle is obtained according to the point cloud of the container side in the 3D point cloud.
- the plane normal vector is obtained.
- the three-dimensional point cloud is converted and converted to the lidar coordinate system.
- the ground point cloud in the three-dimensional point cloud is parallel to the bottom plane of the lidar coordinate system
- the converted container side point cloud is parallel to the side plane of the lidar coordinate system.
- the obtained point cloud data is not affected by the lidar installation angle, installation position, and truck parking position, and the ground point cloud with frontal head-up angle can be obtained.
- the relative translation amount of the spreader and the reference lidar reflects the relative position relationship between the spreader and the reference lidar.
- the relative positional relationship between the spreader and the reference lidar is shown in Figure 5.
- the left side is a three-dimensional view, and the right side is a top view.
- the spreader is in a compressed state.
- D x is the length of the spreader when it is compressed
- D y is the width of the spreader.
- S212 Determine the falling area of the container in the comprehensive point cloud according to the relative translation amount and the size parameter of the container.
- the area below the container can be used as the drop area of the container.
- the container parameters are length X and width Y.
- the spreader picks up a container with length X, the spreader is in an extended state, and the length of the extended part is:
- the container drop area is used as the early warning area. As shown in Figure 8, it is the top view of the RTG.
- the origin 0 is the reference lidar, and the point (T dx , T dy ) is the bottom right corner of the spreader measured in the previous step.
- Detect the coordinate value in the coordinate system X is the container length parameter, Y width represents a value larger than the width of the truck lane, and the reference value is 5.
- z represents the height of the range A of the container drop area. Broadly speaking, the area below the bottom of the container can be counted as an early warning area.
- the value range of Z can be manually set, and the reference value can be set to 6 meters.
- the scope of the drop zone of the container is the area below the container when the spreader clamps the box and packs it down. Therefore, obstacle detection can be carried out in the falling area of the container, and when an obstacle is detected within the falling area of the container, an anti-smashing alarm will be issued.
- obstacles refer to all objects that are not the truck tray within the falling area of the container.
- the above-mentioned three-dimensional laser-based truck anti-smashing detection method has high data source accuracy, and the detection method is not affected by the installation position of the lidar, which greatly improves the accuracy of the anti-smashing detection.
- the lidar includes a reference lidar and at least one alignment lidar installed at a certain angle on the same side under the container crane spreader.
- the reference laser radar collects the first three-dimensional point cloud in the scanning direction, and each is aligned with the second three-dimensional point cloud in the laser radar acquisition and scanning direction.
- the first and second in this embodiment are used to distinguish point clouds collected by different types of lasers. It is understandable that when multiple alignment lidars are provided, the posture angle and position translation of each alignment lidar are used to convert the second three-dimensional point cloud collected separately to the detection system coordinate system.
- the reference lidar 101 is installed on the same side under the spreader 103 of the container crane 102 at a certain angle, and collects the first three-dimensional point cloud of the container operation when the spreader clamps the container and falls.
- the installation position of the lidar is set according to the height of the truck.
- an alignment base light radar 104 is installed at a certain angle, and the alignment lidar 104 collects the second three-dimensional point cloud of the container operation.
- the main control device 105 is in communication connection with the reference laser radar 101 and the alignment-based laser radar 104 respectively.
- the main control device is also connected to the control device 106 of the container crane 102. Both the main control device 105 and the control device 106 can be installed in the control room of the container crane.
- the control device 106 controls the spreader 103 of the container crane 102 to clamp the container 107 in the container stack, and sends the size parameter of the currently clamped container to the main control device 105.
- the control device 106 sends a signal to the main control device 105 that the spreader starts to fall.
- the main control device 105 sends acquisition signals to the reference lidar 101 and the alignment base lidar 104 according to the signal.
- the reference lidar 101 collects the first three-dimensional point cloud
- the alignment lidar 104 collects the second three-dimensional point cloud.
- the lidar includes a reference lidar and an alignment lidar installed at a certain angle on the same side under the container crane spreader;
- the attitude parameters include the first attitude angle of the reference lidar, and the alignment laser radar The second attitude angle of the radar and the positional translation of the alignment lidar relative to the reference lidar.
- Obtaining the three-dimensional point cloud of the container operation collected by the lidar when the container is dropped by the spreader including: obtaining the first three-dimensional point cloud of the container operation collected by the reference lidar when the container is dropped by the spreader and the container, and aligning the lidar The collected second three-dimensional point cloud of container operations.
- the three-dimensional point cloud is converted to obtain the comprehensive point cloud of the container operation, including: converting the first three-dimensional point cloud to the detection system coordinate system according to the first attitude angle; The two-dimensional and three-dimensional point cloud is converted to the inspection system coordinate system; the converted first three-dimensional point cloud and the converted second three-dimensional point cloud are merged to obtain a comprehensive point cloud for container operations.
- the relative translation amount is the relative translation amount of the spreader and the reference lidar.
- the master control device obtains the size parameters of the container currently clamped by the spreader of the container crane during the truck loading operation; obtains the first three-dimensional point cloud collected by the reference lidar when the container is clamped by the spreader and falls, and The second three-dimensional point cloud collected by the quasi-lidar; among them, the reference lidar and the alignment lidar are installed on the same side under the container crane spreader at a certain angle; obtain the first attitude angle of the reference lidar, and obtain the right The second attitude angle and position translation of the quasi-lidar; the first three-dimensional point cloud is converted to the detection system coordinate system according to the first attitude angle; the second three-dimensional point cloud is converted to the detection system according to the second attitude angle and the position translation Coordinate system; fuse the converted first three-dimensional point cloud and converted second three-dimensional point cloud to obtain a comprehensive point cloud for container operations; obtain the relative translation of the spreader and the reference lidar; according to the relative translation and the size of the container Parameters, determine the range of the container's falling area in the
- a three-dimensional laser-based chucking prevention detection method includes the following steps:
- S404 Acquire the first three-dimensional point cloud collected by the reference lidar when the container is clamped by the spreader and fall, and the second three-dimensional point cloud collected by the alignment lidar; wherein, the reference lidar and the alignment lidar are installed on the container crane. The same side under the tool is set at a certain angle.
- Two lidars are used in the truck anti-smashing detection, both of which are installed on the same side under the container crane spreader at a certain angle. Choose one of them as the reference lidar, and the other is the alignment lidar. For example, two lidars are installed on the same side, one in the front, scanning backward at a certain angle to the x-axis; one behind, scanning forward at a certain angle to the x-axis. Therefore, the alignment lidar and the reference lidar can comprehensively obtain the three-dimensional point cloud of the truck-packing operation from two angles.
- the reference laser radar is used as the origin to establish the coordinate system of the entire detection system.
- the origin O represents the position of the reference laser radar
- the X axis is parallel to the container crane arm
- the Y axis direction is perpendicular to the container crane arm
- Z The axis direction is the height direction.
- the cube in the picture represents the position of the container and the truck.
- the control device of the container crane sends a signal that the spreader starts to fall to the main control device, and the main control device sends a signal acquisition signal to the reference lidar and the alignment lidar.
- the reference lidar and The aligning lidar collects 3D point clouds at a set frequency according to the collected signals, and feeds the collected 3D point clouds to the main control device, and the main control device continuously analyzes and judges the falling process of the spreader.
- the reference lidar collects the first three-dimensional point cloud when the container is clamped by the spreader at the current time according to the collected signal, and the second three-dimensional point cloud is collected by the laser radar when the container is clamped by the spreader at the current time and falls.
- S406 Acquire a first attitude angle of the reference lidar, and acquire a second attitude angle and a position translation amount of the alignment lidar.
- the first attitude angle of the reference lidar refers to the installation angle of the reference lidar relative to the reference object, including but not limited to roll angle, pitch angle, and yaw angle.
- the attitude angle of the reference lidar can be determined according to the three-dimensional point cloud of the truck loading operation.
- the attitude angle of the reference lidar only needs to be calculated and stored once, and then the stored first attitude angle can be read for point cloud calibration, or real-time Each test is calibrated sexually, so that the calibrated point cloud will be more accurate.
- the second attitude angle and position translation of the alignment lidar refer to the installation angle and distance of the alignment lidar relative to the detection system coordinate system.
- the second attitude angle includes but is not limited to roll angle, pitch angle and yaw Horn.
- the second attitude angle and position translation of the aligned lidar can be obtained by calibrating the same calibration object with two radars.
- obtaining the first attitude angle of the reference lidar includes the following steps:
- S602 Acquire a first-direction calibration three-dimensional point cloud collected by the reference lidar when the container is clamped and dropped by the spreader in the calibration state.
- the card and boxing operation of the anti-smashing detection using the method of this application for the first time can be regarded as the calibration state.
- the first-direction calibration three-dimensional point cloud is the data collected by the reference lidar when the method of this application is used for the anti-smashing detection for the first time.
- it can also be calibrated at regular intervals.
- the container loading operation of the anti-smashing detection using the method of this application for the first time every week is used as the calibration state.
- the first direction calibration three-dimensional point cloud is every When Zhou first adopted the method of this application for anti-smashing detection, the spreader collected by the reference lidar clamped the first three-dimensional point cloud when the container fell.
- S604 Determine the ground point cloud from the calibration three-dimensional point cloud in the first direction according to the installation height of the reference lidar.
- Ground point cloud refers to the point cloud on the ground determined by the installation position of the reference lidar. It is known that the height of the reference lidar is a, and the point cloud whose z coordinate value is less than -a in the calibration three-dimensional point cloud in the first direction is taken as the ground point cloud.
- the normal vector is a concept of space analytic geometry, and the vector represented by a straight line perpendicular to the plane is the normal vector of the plane.
- the method of calculating the normal vector is to first calculate the covariance matrix of the ground point cloud, and then perform singular value decomposition on the covariance matrix.
- the singular vector obtained by the singular value decomposition describes the three main directions of the point cloud data and is perpendicular to the normal vector of the plane. It represents the direction with the smallest variance, and the smallest variance represents the smallest singular value, so finally the vector with the smallest singular value is selected as the normal vector of the plane.
- C is the covariance matrix
- s i is the point in the point cloud
- S608 Calculate the roll angle and the pitch angle of the reference lidar according to the plane normal vector of the ground point cloud.
- the pitch angle is the angle between the X axis of the reference lidar coordinate system and the horizontal plane
- the roll angle is the angle between the lidar coordinate Y axis and the lidar vertical plane.
- the formula for calculating the roll angle and the pitch angle is:
- T 1 (a 1 ,b 1 ,c 1 )
- T 1 is the normal vector of the ground
- ⁇ is the roll angle
- ⁇ is the pitch angle
- the side point cloud of the container is determined from the calibration three-dimensional point cloud in the first direction.
- the container side point cloud refers to the point cloud representing the side part of the container in the first-direction calibration three-dimensional point cloud collected at the container operation site. It can be determined according to the height of the point cloud and the distance between the point cloud and the lidar.
- the height of the lidar is a
- the height of the truck tray is b
- the height of the container is c.
- the z coordinate range taken is [-a+b,-a+b+c] Point cloud, as a filtered point cloud. Since the side of the container is close to the lidar, a distance threshold t is set, and based on the point cloud after one-time filtering, a point cloud with a distance less than t from the lidar is taken as the side point cloud of the container.
- S612 Calculate the plane normal vector of the point cloud on the side of the container.
- S614 Calculate the yaw angle of the reference lidar according to the plane normal vector of the point cloud on the side of the container.
- the first attitude angle includes the roll angle, the pitch angle, and the yaw angle.
- the yaw angle is the angle between the Z axis of the lidar coordinate system and the side of the container.
- the calculation formula for calculating the yaw angle is:
- T 2 (a 2 ,b 2 ,c 2 )
- T 2 is the plane normal vector of the point cloud on the side of the container
- ⁇ is the yaw angle
- the roll angle, pitch angle and yaw angle of the lidar are calculated by the method of plane normal vector.
- acquiring the second attitude angle of the alignment lidar and the positional translation amount of the alignment lidar relative to the reference lidar includes: acquiring the first three dimensions of the calibration object collected by the reference lidar for the same calibration object Point cloud and the second three-dimensional point cloud of the calibration object collected by the laser radar; convert the first three-dimensional point cloud of the calibration object to the detection system coordinate system; for the second three-dimensional point cloud of the calibration object and the converted first three-dimensional point cloud of the calibration object The point cloud performs point cloud matching to determine the second attitude angle of the alignment lidar and the translation amount of the alignment lidar relative to the reference lidar.
- the first three-dimensional point cloud of the calibration object is converted to the detection system coordinate system, and the first posture angle that has been calibrated is used for conversion.
- the purpose is to calibrate the two lidars so that their point clouds can be converted to the same coordinate system, reducing detection The detection blind zone of the system.
- the first three-dimensional point cloud of the calibration object collected by the reference lidar is converted to the detection system coordinate system, and then the converted calibration object
- the first three-dimensional point cloud and the second three-dimensional point cloud of the calibration object collected by the alignment lidar are used to calculate the posture between the lidars using point cloud matching, and the alignment base light radar is determined according to the difference in point cloud data of the same object in different coordinate systems
- the point cloud matching method can use commonly used icp (Iterative Closest Point), ndt (normal distribution transformation), etc.
- step S406 the method further includes: S408, converting the first three-dimensional point cloud to the detection system coordinate system according to the first attitude angle.
- the coordinate system of the detection system is established with the reference lidar as the origin.
- the first three-dimensional point cloud is transformed to be parallel to the plane of the detection system coordinate system according to the first attitude angle.
- the first three-dimensional point cloud is converted, and the ground point cloud of the converted first three-dimensional point cloud is parallel to the XOY plane of the detection system coordinate system.
- the yaw angle transforms the converted first three-dimensional point cloud, and the container side point cloud of the converted first three-dimensional point cloud is parallel to the XOZ of the coordinate system of the detection system.
- the first three-dimensional point cloud is rotated around the X axis of the detection system coordinate system
- the first three-dimensional point cloud is rotated around the Y axis of the detection system coordinate system
- Convert the ground point cloud in the first three-dimensional point cloud to be parallel to the bottom plane of the lidar coordinate system.
- R x and R y are rotation matrices around the x-axis and around the y-axis
- p g is the ground point cloud in the first three-dimensional point cloud parallel to the XOY plane of the detection system coordinate system after conversion
- p c is the original ground point cloud.
- the converted first three-dimensional point cloud is rotated around the Z axis of the detection system coordinate system, and the converted first three-dimensional point cloud is between the container side point cloud and the lidar coordinate system.
- the side planes are parallel.
- R z is the rotation matrix around the z axis
- p g is the point cloud parallel to the ground and the XOY plane after conversion
- p is the point cloud parallel to the XOZ plane of the detection system coordinate system after the final conversion.
- S410 Convert the second three-dimensional point cloud to the detection system coordinate system according to the second attitude angle and the position translation amount.
- the rotation matrix of the alignment lidar relative to the reference lidar that has been converted to the detection system coordinate system is determined according to the second attitude angle and the position translation amount, and the second three-dimensional point cloud is converted to the detection system coordinate system according to the rotation matrix.
- the planes are parallel.
- p l is the original second three-dimensional point cloud collected by the laser radar
- p lg is the second three-dimensional point cloud converted in the coordinate system of the detection system.
- the converted first three-dimensional point cloud and the converted second three-dimensional point cloud are merged to obtain a comprehensive point cloud for container operations.
- the first three-dimensional point cloud of the reference laser radar in the detection system coordinate system is p g
- the relative translation amount of the spreader and the reference lidar reflects the relative position relationship between the spreader and the reference lidar.
- the relative positional relationship between the spreader and the reference lidar is shown in Figure 5.
- the left side is a three-dimensional view, and the right side is a top view.
- the spreader is in a compressed state.
- D x is the length of the spreader when it is compressed
- D y is the width of the spreader.
- S416 Determine the falling area of the container in the comprehensive point cloud according to the relative translation amount and the size parameter of the container.
- the area below the container can be used as the drop area of the container.
- the drop area of the container is the area below the container when the spreader clamps the box and falls. Therefore, obstacle detection can be performed in the drop area of the container, and when an obstacle is detected within the drop area of the container, an anti-smashing alarm will be issued.
- obstacles refer to all objects other than the truck tray within the falling area of the container.
- the steps of issuing an anti-smashing alarm include:
- S1102 Filter out the truck carrier point cloud from the comprehensive point cloud according to the height threshold of the truck carrier.
- the fluctuation range of the truck tray height range is not large, and the truck tray height threshold H h can be set according to empirical values.
- the truck tray point cloud is the point cloud whose Z coordinate value is less than the height threshold in the comprehensive point cloud.
- the through filtering is performed on the full point cloud p R to remove the truck tray area.
- the side view of the truck and its warning area and height threshold are shown in Figure 14.
- S1104 Perform filtering processing on the point cloud within the drop area of the container.
- the detection system After receiving the signal from the control device that the spreader starts to fall, the detection system performs point cloud denoising filtering processing on the point cloud p X within the falling area of the container.
- the adopted denoising filtering algorithm is radius point filtering, which is filtered according to the number of adjacent points in the radius of the space point. Only if there are point clouds greater than the set threshold in a certain range, they are retained.
- S1106 Perform obstacle detection on the point cloud within the falling area of the container after the filtering process.
- the obstacle classification algorithm can be used to detect.
- the main control device If it detects that there is an obstacle point cloud, it will be determined that there is a possibility of smashing, and the main control device outputs an anti-smashing alarm signal to the control equipment of the container crane to give an anti-smashing warning.
- obstacle detection is performed within the falling area of the container by filtering out the bracket point cloud.
- an anti-smashing alarm when an obstacle is detected within the falling area of the container, an anti-smashing alarm is issued, including:
- each comprehensive point cloud it is expressed in pixels to obtain a two-dimensional image.
- the steps of projecting a full point cloud into a two-dimensional image include:
- the coordinates of its two-dimensional image can be calculated by the following formula.
- u and v are the row and column coordinates of the two-dimensional image
- x i and z i are the x-axis and z-axis coordinates of the i-th comprehensive point cloud
- x min and z min are the x-axis and z-axis coordinates of the comprehensive point cloud.
- the minimum value of the Z axis, u r and v r are the accuracy of the comprehensive point cloud projected onto the two-dimensional image, which represents the actual distance between adjacent pixels on the two-dimensional image.
- pixels are used to represent the comprehensive point cloud
- the coordinates of the pixel points are the two-dimensional coordinates of the comprehensive point cloud.
- S1406 Binarize point cloud pixels and non-point cloud pixels to obtain a binary image.
- the binarization process refers to the process of setting the gray value of the pixel on the image to 0 or 255, that is, the process of presenting the entire image with a clear black and white effect.
- One way may be to set the gray value of the pixel converted from the point cloud to 255, and set the gray value of other non-point cloud converted pixels to 0 to obtain a binary image.
- Another way can be to set the gray value of the pixel converted from the point cloud to 0, and set the gray value of other non-point cloud converted pixels to 255 to obtain a binary image.
- S1408 Perform image preprocessing on the binary image to obtain a two-dimensional image.
- the image preprocessing includes: firstly perform median filtering and bilateral filtering preprocessing operations on the two-dimensional image.
- the median filtering is to protect the edge information
- the bilateral filtering is to preserve the edges and denoise; and then perform the morphological expansion operation. Due to the scanning method of the laser sensor, the distance between some adjacent points will be greater than the pixel distance of the image, resulting in holes in the image. If the pixel accuracy is increased, the resolution of the image will be reduced.
- the expansion operation on the image can effectively reduce Hole.
- Image preprocessing methods are not limited to morphological expansion. It is also possible to perform morphological closing operations on the image to fill the black hole area, and then perform morphological opening operations to enhance edge information and filter discrete interference pixels.
- step S1302 it further includes:
- S1304 Determine the range of the drop area of the container and the position range of the truck tray in the two-dimensional image.
- S1308 Perform image detection on the falling area of the container in the two-dimensional image. If an obstacle is detected, an anti-smashing alarm is issued.
- the image detection is performed on the pixels in the falling area of the container in the two-dimensional image to obtain the detection result of the anti-smashing of the truck.
- image detection is performed on the pixels in the location range of the container falling area in the two-dimensional image, and if an obstacle is detected, the steps of issuing an anti-smashing alarm include:
- S1702 Traverse each row within the location range of the container drop area in the two-dimensional image, and count the number of point cloud pixels in each row.
- the point cloud pixel refers to the pixel converted from the point cloud.
- the gray value of point cloud pixels can be 255
- the gray value of non-point cloud pixels can be 0.
- the gray value of the point cloud pixel can be 0, and the gray value of the non-point cloud pixel is 255.
- the gray values of the point cloud pixels the number of pixels in each row within the position range of the container drop area in the two-dimensional image whose gray values are corresponding values is counted.
- the gray value of a point cloud pixel is 255
- count the number of pixels with a gray value of 255 in each row of the two-dimensional image that is, count how many pixels in each row have a pixel value of 255. In this way, the number of point cloud pixels in each row is obtained.
- step S1706 is executed, and if the number of point cloud pixels in the current line is less than the first threshold, step S1708 is executed.
- the counter is increased by a preset value.
- Step S1708 is executed after step S1706.
- S1708 Determine whether the traversal of each row within the location range of the container drop area is completed.
- step S1710 If yes, go to step S1710, if no, go back to step S1702.
- step S1712 is executed.
- the position range of the drop area of the container does not collect a three-dimensional point cloud.
- the number of point cloud pixels in each row in the drop area of the container is zero.
- the 3D point cloud is collected from the location range of the drop area of the container.
- the number of pixels in the rows is greater than 0, and the number of rows exceeds the threshold T1, that is, if the statistical value of the counter is greater than T2, the obstacle can be detected.
- the first threshold and the second threshold can be set according to accuracy requirements and empirical values.
- This method can be applied to the container gantry crane equipment at the container terminal.
- the system can judge the possibility of the truck head or other container on the pallet being hit by the falling container, and avoid accidents. It can adapt to the condition of 20ft, double 20ft, 40ft and 45ft container with cassette.
- a three-dimensional laser-based truck anti-smashing detection device includes: a container acquisition module 2002 for acquiring the current spreader of a container crane in a truck loading operation. The size parameters of the clamped container; the point cloud acquisition module 2004, which acquires the three-dimensional point cloud of the container operation collected by the lidar when the container is clamped by the spreader, and the attitude parameter acquisition module 2006, which is used to acquire the attitude parameters of the lidar.
- the conversion module 2008 is used to convert the three-dimensional point cloud according to the attitude parameters to obtain the comprehensive point cloud of the container operation; the position acquisition module 2010 is used to obtain the relative translation between the spreader and the lidar; the falling area determination module 2012 is used According to the relative translation amount and the size parameters of the container, determine the range of the container drop area in the comprehensive point cloud; the detection module 2014 is used to issue an anti-smashing alarm when an obstacle is detected within the range of the container drop area.
- the lidar includes a reference lidar and an alignment lidar installed at a certain angle on the same side under the container crane spreader;
- the attitude parameters include the first attitude angle of the reference lidar, and the alignment laser radar The second attitude angle of the radar and the positional translation of the alignment lidar relative to the reference lidar.
- the point cloud acquisition module is used to acquire the first three-dimensional point cloud of the container operation collected by the reference laser radar when the container is dropped by the spreader, and the second three-dimensional point cloud of the container operation collected by the laser radar.
- the conversion module includes: a first conversion module for converting the first three-dimensional point cloud to the detection system coordinate system according to the first attitude angle; a second conversion module for converting the second three-dimensional point cloud according to the second attitude angle and position translation
- the point cloud is converted to the inspection system coordinate system; the fusion module is used to fuse the converted first three-dimensional point cloud and the converted second three-dimensional point cloud to obtain a comprehensive point cloud for container operations; the position acquisition module is used to obtain the spreader The amount of relative translation from the reference lidar.
- the attitude parameter acquisition module includes a calibration point cloud acquisition module, which is used to acquire the first direction calibration three-dimensional point cloud collected by the reference lidar when the spreader clamps the container and falls in the calibration state.
- the ground point cloud determination module is used to determine the ground point cloud from the calibration three-dimensional point cloud in the first direction according to the installation height of the reference lidar;
- the normal vector calculation module is used to calculate the plane normal vector of the ground point cloud;
- the angle determination module It is used to calculate the roll angle and pitch angle of the reference lidar according to the plane normal vector of the ground point cloud;
- the side point cloud determination module is used to calculate the installation height of the reference lidar, the height of the truck tray, the height of the container and the reference laser
- the distance of the radar is used to determine the point cloud of the side of the container from the calibration three-dimensional point cloud in the first direction;
- the normal vector calculation module is also used to calculate the plane normal vector of the point cloud of the side of the container;
- the angle determination module is also used to determine the point cloud
- the attitude parameter acquisition module further includes: a calibration module for acquiring the first three-dimensional point cloud of the calibration object collected by the Lidar for the same calibration object, and the second three-dimensional point cloud of the calibration object collected by the reference Lidar Point cloud; the first conversion module is also used to convert the first three-dimensional point cloud of the calibration object to the detection system coordinate system; the matching module is used to convert the converted first three-dimensional point cloud of the calibration object and the second three-dimensional point cloud of the calibration object Perform point cloud matching to determine the second attitude angle and position translation of the laser radar.
- the detection module includes: a point cloud filtering module, which is used to filter out the point cloud of the truck carrier from the comprehensive point cloud according to the height threshold of the truck carrier; and the filtering processing module is used to filter the container The point cloud within the falling area is filtered; the obstacle detection module is used to perform obstacle detection on the point cloud within the falling area of the container after the filtering process; the alarm module is used to detect obstacles if the point cloud is detected. An anti-smashing alarm is issued.
- a point cloud filtering module which is used to filter out the point cloud of the truck carrier from the comprehensive point cloud according to the height threshold of the truck carrier
- the filtering processing module is used to filter the container The point cloud within the falling area is filtered
- the obstacle detection module is used to perform obstacle detection on the point cloud within the falling area of the container after the filtering process
- the alarm module is used to detect obstacles if the point cloud is detected. An anti-smashing alarm is issued.
- the detection module further includes: a projection module, which is used to project the full point cloud into a two-dimensional image; Position range; pixel point filtering module, used to remove the pixel points of the truck tray in the two-dimensional image according to the determined position range of the truck tray in the two-dimensional image; pixel point detection module, used for the two-dimensional In the image, the pixels in the falling area of the container are detected. If an obstacle is detected, an anti-smashing alarm will be issued.
- a projection module which is used to project the full point cloud into a two-dimensional image
- Position range Position range
- pixel point filtering module used to remove the pixel points of the truck tray in the two-dimensional image according to the determined position range of the truck tray in the two-dimensional image
- pixel point detection module used for the two-dimensional In the image, the pixels in the falling area of the container are detected. If an obstacle is detected, an anti-smashing alarm will be issued.
- the projection module includes: a coordinate calculation module for calculating the two-dimensional coordinates of each comprehensive point cloud; a pixel point conversion module for calculating the two-dimensional coordinates of each comprehensive point cloud, Convert the point cloud into pixels; the binarization module is used to binarize point cloud pixels and non-point cloud pixels to obtain a binary image; the preprocessing module is used to perform image preprocessing on the binary image Processing to obtain a two-dimensional image.
- the pixel point detection module includes: a traversal module, which is used to traverse each row within the position range of the container drop area in the two-dimensional image, and count the number of point cloud pixels in each row; a counter is used to determine the current row If the number of point cloud pixels is greater than the first threshold, the counter is increased by a preset value; the comparison module is used to compare the statistical value of the counter with the second threshold after the traversal of each line within the location range of the container drop area is completed; The smashing detection module is used to obtain the detection result of the detected obstacle if the statistical value of the counter is greater than the second threshold value, and issue an anti-smashing alarm.
- the various modules in the above-mentioned three-dimensional laser-based truck anti-smashing detection device can be implemented in whole or in part by software, hardware, and a combination thereof.
- the above-mentioned modules may be embedded in the form of hardware or independent of the processor in the computer equipment, or may be stored in the memory of the computer equipment in the form of software, so that the processor can call and execute the operations corresponding to the above-mentioned modules.
- a computer device is provided.
- the computer device may be the main control device in FIG. 1, and its internal structure diagram may be as shown in FIG. 21.
- the computer equipment includes a processor, a memory, a communication interface, a display screen and an input device connected through a system bus. Among them, the processor of the computer device is used to provide calculation and control capabilities.
- the memory of the computer device includes a non-volatile storage medium and an internal memory.
- the non-volatile storage medium stores an operating system and a computer program.
- the internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium.
- the communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be implemented through WIFI, an operator's network, NFC (near field communication) or other technologies.
- the computer program is executed by the processor to realize a three-dimensional laser-based truck anti-smashing detection method.
- the display screen of the computer equipment can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer equipment can be a touch layer covered on the display screen, or it can be a button, trackball or touch pad set on the housing of the computer equipment , It can also be an external keyboard, touchpad, or mouse.
- FIG. 21 is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied.
- a computer device including a memory and a processor, and a computer program is stored in the memory.
- the processor executes the computer program, the steps of the three-dimensional laser collection card anti-smashing detection method of the foregoing embodiments are implemented. .
- a computer-readable storage medium is provided, and a computer program is stored thereon.
- the computer program is executed by a processor, the steps of the three-dimensional laser collection card anti-smashing detection method of the above-mentioned embodiments are implemented.
- Non-volatile memory may include read-only memory (Read-Only Memory, ROM), magnetic tape, floppy disk, flash memory, or optical storage.
- Volatile memory may include random access memory (RAM) or external cache memory.
- RAM may be in various forms, such as static random access memory (Static Random Access Memory, SRAM) or dynamic random access memory (Dynamic Random Access Memory, DRAM), etc.
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Mechanical Engineering (AREA)
- Computer Networks & Wireless Communication (AREA)
- Electromagnetism (AREA)
- General Physics & Mathematics (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Automation & Control Theory (AREA)
- Aviation & Aerospace Engineering (AREA)
- Length Measuring Devices By Optical Means (AREA)
Abstract
一种基于三维激光的集卡防砸检测方法、装置和计算机设备。该方法包括:获取集卡装箱作业中集装箱起重机的吊具当前所夹住的集装箱的尺寸参数(S202);获取在吊具夹住集装箱下落时激光雷达采集的集装箱作业的三维点云(S204);获取激光雷达的姿态参数(S206);根据姿态参数,对三维点云进行转换,得到集装箱作业的全面点云(S208);获取吊具与激光雷达的相对平移量(S210);根据相对平移量以及集装箱的尺寸参数,在全面点云中确定集装箱下落区域范围(S212);当在集装箱下落区域范围内检测到障碍物时,发出防砸警报(S214)。该方法的数据源精度高,且检测方法不受激光雷达安装位置的影响,极大地提高了防砸检测的精度。
Description
本申请涉及激光雷达技术领域,特别是涉及一种基于三维激光的集卡防砸检测方法、装置和计算机设备。
集卡装箱作业是指利用集装箱起重机的吊具夹住处于集装箱垛的集装箱,将其吊起,并控制其下落装裁至集卡的托架上。
在集装箱门式起重机进行集卡装箱作业时,由于集卡未停放到位,下落集装箱将会与集卡车头或托架上其他集装箱发生碰砸。为解决这一问题,集卡装箱作用引入了二维激光扫描仪进行检测,二维激光扫描仪沿平行于集卡车道中心线对集装箱及集卡进行扫描测距,实时计算集卡偏离场桥起吊点的距离值,通过led屏幕告知司机调整集卡位置,从而达到防砸的目的。
由于使用的是二维激光雷达,使得该方法只能对集卡在前后两个方向纠偏调整,且纠正的方向受激光雷达的安装位置的影响,从而导致防砸检测精度低。
发明内容
基于此,有必要针对上述技术问题,提供一种能够提高检测精度的基于三维激光的集卡防砸检测方法、装置和计算机设备。
一种基于三维激光的集卡防砸检测方法,所述方法包括:
获取集卡装箱作业中集装箱起重机的吊具当前所夹住的集装箱的尺寸参数;
获取在所述吊具夹住所述集装箱下落时激光雷达采集的集装箱作业的三维点云;
获取所述激光雷达的姿态参数;
根据所述姿态参数,对所述三维点云进行转换,得到集装箱作业的全面点云;
获取吊具与所述激光雷达的相对平移量;
根据所述相对平移量以及所述集装箱的尺寸参数,在所述全面点云中确定 集装箱下落区域范围;
当在所述集装箱下落区域范围内检测到障碍物时,发出防砸警报。
一种基于三维激光的集卡防砸检测装置,所述装置包括:
集装箱获取模块,用于获取集卡装箱作业中集装箱起重机的吊具当前所夹住的集装箱的尺寸参数;
点云获取模块,获取在所述吊具夹住所述集装箱下落时激光雷达采集的集装箱作业的三维点云;
姿态参数获取模块,用于获取所述激光雷达的姿态参数;
转换模块,用于根据所述姿态参数,对所述三维点云进行转换,得到集装箱作业的全面点云;
位置获取模块,用于获取吊具与所述基准激光雷达的相对平移量;
下落区域确定模块,用于根据所述相对平移量以及所述集装箱的尺寸参数,在所述全面点云中确定集装箱下落区域范围;
检测模块,用于当在所述集装箱下落区域范围内检测到障碍物时,发出防砸警报。
一种计算机设备,包括存储器和处理器,所述存储器存储有计算机程序,所述处理器执行所述计算机程序时实现上述各实施例的方法的步骤。
一种计算机可读存储介质,其上存储有计算机程序,所述计算机程序被处理器执行时实现以下上述各实施的方法的步骤。
上述基于三维激光的集卡防砸检测方法、装置、计算机设备和存储介质,利用激光雷达采集吊具夹住集装箱下落时集装箱作业的三维数据,数据精度高,在高精度三维点云数据的基础上,根据姿态参数对三维点云进行转换得到全面点云,从而监控范围不受激光雷达安装位置的影响,进而根据夹具所夹住的集装箱尺寸参数,吊具与基准激光雷达的平移量,在全面点云中确定集装箱下落区域范围,当在集装箱下落区域范围内检测到障碍物时,发出防砸警报。该方法的数据源精度高,且检测方法不受激光雷达安装位置的影响,极大地提高了防砸检测的精度。
图1为一个实施例中基于三维激光的集卡防砸检测方法的应用环境图;
图2为一个实施例中集装箱装箱作业中集装箱下落区域无障碍物的场景示意图;
图3为一个实施例中集装箱装箱作业中集装箱下落区域有障碍物的场景示意图;
图4为一个实施例中基于三维激光的集卡防砸检测方法的流程示意图;
图5为一个实施例中吊具与基准激光雷达的相对位置关系示意图;
图6为另一个实施例中吊具与基准激光雷达的相对位置关系示意图;
图7为一个实施例中吊具吊起集装箱时吊具与基准激光雷达的相对位置关系示意图;
图8为一个实施例中集装箱下落区域的示意图;
图9为另一个实施例中基于三维激光的集卡防砸检测方法的应用环境图;
图10为另一个实施例中基于三维激光的集卡防砸检测方法的流程示意图;
图11为一个实施例中检测系统坐标系的设定示意图;
图12为一个实施例中获取基准激光雷达的第一姿态角步骤的流程示意图;
图13为一个实施例中当在集装箱下落区域范围内检测到障碍物时,发出防砸警报的步骤的流程示意图;
图14为一个实施例中集卡侧视图及其预警区域、高度阈值的关系示意图;
图15为另一个实施例中当在集装箱下落区域范围内检测到障碍物时,发出防砸警报的步骤的流程示意图;
图16为一个实施例中将全面点云投影为二维图像的步骤的流程示意图;
图17为一个实施例中集装箱下方无障碍物的二维图像示意图;
图18为一个实施例中集装箱下方有障碍物的二维图像示意图;
图19为一个实施例中对二维图像中的集装箱下落区域的位置范围的像素点进行图像检测,若检测到障碍物,发出防砸警报的步骤的流程示意图;
图20为一个实施例中基于三维激光的集卡防砸检测装置的结构框图;
图21为一个实施例中计算机设备的内部结构图。
为了使本申请的目的、技术方案及优点更加清楚明白,以下结合附图及实施例,对本申请进行进一步详细说明。应当理解,此处描述的具体实施例仅仅用以解释本申请,并不用于限定本申请。
本申请提供的基于三维激光的集卡防砸检测方法,可以应用于如图1所示的应用环境中。其中,激光雷达101安装在集装箱起重机102的吊具103的下方的同一侧呈一定夹角,采集吊具夹住集装箱下落时集装箱作业的三维点云。激光雷达的安装位置根据集卡高度设置。主控设备105与激光雷达101通信连接。主控设备还与集装箱起重机102的控制设备106连接。主控设备105以及控制设备106均可设置在集装箱起重机的控制机房内。
在集卡进行装箱作业时,控制设备106控制集装箱起重机102的吊具103夹住处于集装箱堆垛中的集装箱107时,向主控设备105发送当前所夹住的集装箱的尺寸参数。当吊具作业吊起该集装箱时准备装载至集卡108时,控制设备106向主控设备105发送吊具开始下落的信号。主控设备105根据该信号向激光雷达101发送采集信号,激光雷达101采集集装箱作业的三维点云。
如图2所示,集装箱下落时,集装箱正下方的集装箱下落区域,无障碍物的物体,判定不会发生碰砸。如图3所示,集装箱下落,集装箱正下方的集装箱下落区域,激光雷达检测到有一定高度的障碍物(如集卡车头,集卡托架上的其它集装箱等),输出防砸警报信号给集装箱起重机的控制设备。
如图4所示,提供一种基于三维激光的集卡防砸检测方法,以该方法应用于图1中的主控设备为例进行说明,包括以下步骤:
S202,获取集卡装箱作业中集装箱起重机的吊具当前所夹住的集装箱的尺寸参数。
集卡装箱作业是指利用集装箱起重机的吊具夹住处于集装箱垛的集装箱,将其吊起,并控制其下落装裁至集卡的托架上。当吊具作业夹住处于集装箱堆垛中的集装箱时,集装箱起重机的控制设备发送此时吊具夹住的集装箱尺寸参数给主控设备。其中,集装箱的尺寸参数包括集装箱的长、宽和高。
S204,获取在吊具夹住集装箱下落时激光雷达采集的集装箱作业的三维点云。
具体地,激达雷达采集集装箱作业现场的三维激光点云。当吊具作业吊起该集装箱时准备装载至集卡时,控制设备向主控设备发送吊具开始下落的信号。主控设备根据该信号向激光雷达发送采集信号,激光雷达采集集装箱作业的三维点云。
S206,获取激光雷达的姿态参数。
其中,姿态参数包括姿态角,是指激光雷达相对参照物的安装角度,包括但不限于翻滚角、俯仰角和偏航角。激光雷达的姿态角可根据集装箱作业的三维点云确定。在实际应用中,由于集卡防砸检测装置安装完成后,激光雷达的位置基本固定,激光雷达的姿态角,只需要计算一次,后续可以使用第一次的姿态角,进行点云校准,也可以实时性地对每一次检测都进行校准,这样校准的点云会更加精确。
一个实施例中,姿态角包括翻滚角、俯仰角和偏航角。具体地,获取激光雷达的姿态角的步骤包括:获取在标定状态激光雷达采集的集装箱作业的三维标定点云;根据激光雷达的安装高度,从三维标定点云中确定地面点云;计算地面点云的平面法向量;根据地面点云的平面法向量,计算激光雷达的翻滚角和俯仰角;根据激光雷达的安装高度、集卡托架高度、集装箱高度以及与激光雷达的距离,从集装箱作业三维点云中确定集装箱侧面点云;计算集装箱侧面点云的平面法向量;根据集装箱侧面点云的平面法向量,计算激光雷达的偏航角。
S208,根据姿态参数,对三维点云进行转换,得到集装箱作业的全面点云。
姿态角包括了翻滚角、俯仰角和偏航角,其中,翻滚角和俯仰角根据三维点云中的地面点云的平面法向量得到,偏航角根据三维点云中的集装箱侧面点云的平面法向量得到。具体地,根据姿态参数,对三维点云进行转换,转换至激光雷达坐标系。具体地,转换后,三维点云中的地面点云与激光雷达坐标系的底平面平行,转换后的集装箱侧面点云与激光雷达坐标系的侧平面平行。通过转换后,得到的点云数据不受激光雷达安装角度、安装位置以及集卡停放位置的影响,能够得到正面平视角度的地面点云。
S210,获取吊具与激光雷达的相对平移量。
具体地,吊具与基准激光雷达的相对平移量,反应了吊具与基准激光雷达的相对位置关系。本实施例中,手动测量吊具在缩紧状态,即位于最高位置时吊具最右下角到基准激光雷达的相对平移量T
d=(T
dx T
dy 0)。吊具与基准激光雷达的相对位置关系如图5所示,左侧为三维图,右侧为俯视图,此时吊具处于紧缩状态。如图6所示,D
x为吊具紧缩时的长,D
y为吊具的宽。
S212,根据相对平移量以及集装箱的尺寸参数,在全面点云中确定集装箱下落区域范围。
在实用应用中,集装箱以下的区域,都可以作为集装箱下落区域范围。
具体地,如图7所示,集装箱参数为长为X,宽为Y,当吊具抓起长为X的集装箱时,吊具处于伸长状态,伸长部分长度为:
令吊具在伸长状态与地面的距离(表示吊具可继续伸长的长度)为ΔX,对上式变换,则可得到:
则集装箱下落区域范围为:
A={x,y+T
dy,z}
集装箱下落区域作为预警区域,如图8所示,为整个图为RTG的俯视图,原点0为基准激光雷达,点(T
dx,T
dy)为上一步骤中测量得到的吊具最右下角在检测坐标系中的坐标值,X为集装箱长度参数,Y
宽表示的是大于集卡作业车道宽度的一个值,参考值为5。z表示集装箱下落区域范围A的高度,广义上来说,集装箱底面以下的区域,都可算进预警区域。但由于激光雷达安装位置限制,当集装箱位于高位(如>6米)时,可能无法检测得到集装箱的实际高度,所以可以人 为设置Z的取值范围,参考值可设为6米。
S214,当在集装箱下落区域范围内检测到障碍物时,发出防砸警报。
如前面所提及的,集装箱下落区域范围,是吊具夹住箱装箱下落时集装箱以下的区域。因此,可将在箱装箱下落区域进行障碍物检测,当在集装箱下落区域范围内检测到障碍物时,发出防砸警报。其中,障碍物是指集装箱下落区域范围内非集卡托架的一切物体。
上述基于三维激光的集卡防砸检测方法,数据源精度高,且检测方法不受激光雷达安装位置的影响,极大地提高了防砸检测的精度。
在实际应用中,对于超长的集卡车,如双20ft集装箱,受激光雷达扫描范围的影响,可能存在盲区,而导致防砸检测不准确。为避免盲区,提高检测准确度,激光雷达包括安装在集装箱起重机吊具下方的同一侧呈一定夹角设置的基准激光雷达和至少一个对准激光雷达。基准激光雷达采集扫描方向的第一三维点云,各对准激光雷达采集扫描方向的第二三维点云。本实施例中的第一和第二用于区分不同类型激光采集的点云。可以理解的是,当设置有多个对准激光雷达时,利用各对准激光雷达的姿态角和位置平移量,将各自采集的第二三维点云转换至检测系统坐标系。
本实施例以采用两个激光雷达进行扫描为例进行说明。如图9所示,基准激光雷达101安装在集装箱起重机102的吊具103的下方的同一侧呈一定夹角,采集吊具夹住集装箱下落时集装箱作业的第一三维点云。激光雷达的安装位置根据集卡高度设置。激光雷达101的同侧呈一定角度安装有对准基光雷达104,对准激光雷达104采集集装箱作业的第二三维点云。其中,由于基准激光雷达和对准激光雷达位于吊具下方的同一侧呈一定夹角,因此,利用两个激光雷达采集的第一三维点云和第二三维点云综合为全面的集装箱作业三维点云,无死角扫描集装箱作业场景三维数据。主控设备105分别与基准激光雷达101、对准基光雷达104通信连接。主控设备还与集装箱起重机102的控制设备106连接。主控设备105以及控制设备106均可设置在集装箱起重机的控制机房内。
在集卡进行装箱作业时,控制设备106控制集装箱起重机102的吊具103夹住处于集装箱堆垛中的集装箱107时,向主控设备105发送当前所夹住的集装 箱的尺寸参数。当吊具作业吊起该集装箱时准备装载至集卡108时,控制设备106向主控设备105发送吊具开始下落的信号。主控设备105根据该信号向基准激光雷达101和对准基光雷达104发送采集信号,基准激光雷达101采集第一三维点云,对准激光雷达104采集第二三维点云。
即,本实施例中,激光雷达包括安装在集装箱起重机吊具下方的同一侧呈一定夹角设置的基准激光雷达和对准激光雷达;姿态参数包括基准激光雷达的第一姿态角,对准激光雷达的第二姿态角和对准激光雷达相对于基准激光雷达的位置平移量。
获取在吊具夹住集装箱下落时激光雷达采集的集装箱作业的三维点云,包括:获取在吊具夹住集装箱下落时基准激光雷达采集的集装箱作业的第一三维点云,以及对准激光雷达采集的集装箱作业的第二三维点云。
根据姿态参数,对三维点云进行转换,得到集装箱作业的全面点云,包括:根据第一姿态角将第一三维点云转换至检测系统坐标系;根据第二姿态角和位置平移量将第二三维点云转换至检测系统坐标系;融合转换后的第一三维点云和转换后的第二三维点云,得到集装箱作业的全面点云。
其中,相对平移量为吊具与基准激光雷达的相对平移量。
具体地,主控设备获取集卡装箱作业中集装箱起重机的吊具当前所夹住的集装箱的尺寸参数;获取在吊具夹住集装箱下落时基准激光雷达采集的第一三维点云,以及对准激光雷达采集的第二三维点云;其中,基准激光雷达和对准激光雷达安装在集装箱起重机吊具下方的同一侧呈一定夹角设置;获取基准激光雷达的第一姿态角,以及获取对准激光雷达的第二姿态角和位置平移量;根据第一姿态角将第一三维点云转换至检测系统坐标系;根据第二姿态角和位置平移量将第二三维点云转换至检测系统坐标系;融合转换后的第一三维点云和转换后的第二三维点云,得到集装箱作业的全面点云;获取吊具与基准激光雷达的相对平移量;根据相对平移量以及集装箱的尺寸参数,在全面点云中确定集装箱下落区域范围;当在集装箱下落区域范围内检测到障碍物时,发出防砸警报。
具体地,如图10所示,一种基于三维激光的集卡防砸检测方法,包括以下步骤:
S402,获取集卡装箱作业中集装箱起重机的吊具当前所夹住的集装箱的尺寸参数。
S404,获取在吊具夹住集装箱下落时基准激光雷达采集的第一三维点云,以及对准激光雷达采集的第二三维点云;其中,基准激光雷达和对准激光雷达安装在集装箱起重机吊具下方的同一侧呈一定夹角设置。
集卡防砸检测用到了两个激光雷达,均安装在集装箱起重机吊具下方的同一侧呈一定夹角设置,任选其中一个为基准激光雷达,则另一个为对准激光雷达。例如,两个激光雷达安装在同一侧,一个在前,与x轴成一定角度向后扫描;一个在后,与x轴成一定角度向前扫描。从而,对准激光雷达和基准激光雷达能够从两个角度全面获取集卡装箱作业的三维点云。
其中,如图11以基准激光雷达为原点,建立整个检测系统的坐标系,原点O代表了基准激光雷达的位置,X轴平行于集装箱起重机车臂,Y轴方向垂直于集装箱起重机车臂,Z轴方向为高度方向。图中正方体代表了集装箱和集卡位置。
当集装箱起重机的吊具作业吊起集装箱时,集装箱起重机的控制设备发送吊具开始下落的信号给主控设备,主控设备向基准激光雷达和对准激光雷达发信采集信号,基准激光雷达和对准激光雷达均根据采集信号,以设定频率采集三维点云,并将采集的三维点云反馈给主控设备,由主控设备对在吊具下落过程中持续进行分析判断。具体地,基准激光雷达根据采集信号采集当前时刻在吊具夹住集装箱下落时的第一三维点云,对准激光雷达采集当前时刻在吊具夹住集装箱下落时的第二三维点云。
S406,获取基准激光雷达的第一姿态角,以及获取对准激光雷达的第二姿态角和位置平移量。
其中,基准激光雷达的第一姿态角,是指基准激光雷达相对参照物的安装角度,包括但不限于翻滚角、俯仰角和偏航角。基准激光雷达的姿态角可根据集卡装箱作业的三维点云确定。
在实际应用中,由于激光雷达安装的位置基本固定,基准激光雷达的姿态角,只需要计算一次并存储,后续可以读取已存储的第一次的姿态角,进行点云校准,也可以实时性地对每一次检测都进行校准,这样校准的点云会更加精确。
其中,对准激光雷达的第二姿态角和位置平移量,是指对准激光雷达相对于检测系统坐标系的安装角度及距离,第二姿态角包括但不限于翻滚角、俯仰角和偏航角。对准激光雷达的第二姿态角和位置平移量可根据两个雷达对同一个标定物体进行标定得到。
在一个实施例中,如图12所,获取基准激光雷达的第一姿态角,包括以下步骤:
S602,获取在标定状态吊具夹住集装箱下落时基准激光雷达采集的第一方向标定三维点云。
其中,可以将首次采用本申请方法进行防砸检测的集卡装箱作业作为标定状态,此时,第一方向标定三维点云为首次采用本申请方法进行防砸检测时,基准激光雷达采集的吊具夹住集装箱下落时的第一三维点云。为确保姿态角数据的准确性,还可以定时进行标定,如将每周首次采用本申请方法进行防砸检测的集卡装箱作业作为标定状态,此时,第一方向标定三维点云为每周首次采用本申请方法进行防砸检测时,基准激光雷达采集的吊具夹住集装箱下落时的第一三维点云。
S604,根据基准激光雷达的安装高度,从第一方向标定三维点云中确定地面点云。
地面点云,是指通过基准激光雷达安装位置确定的位于地面的点云。已知基准激光雷达的高度为a,取第一方向标定三维点云中z坐标值小于-a的点云,作为地面点云。
S606,计算地面点云的平面法向量。
法向量,是空间解析几何的一个概念,垂直于平面的直线所表示的向量为该平面的法向量。
计算法向量的方法,首先计算地面点云的协方差矩阵,然后对协方差矩阵进行奇异值分解,奇异值分解得到的奇异向量描述了点云数据的三个主要方向,垂直于平面的法向量代表了方差最小的方向,方差最小代表了奇异值最小,所以最后选取奇异值最小的向量作为平面的法向量。
S608,根据地面点云的平面法向量,计算基准激光雷达的翻滚角和俯仰角。
其中,俯仰角为基准激光雷达坐标系X轴与水平面的夹角,翻滚角为激光雷达坐标Y轴与激光雷达铅垂面的夹角。
具体地,计算翻滚角和俯仰角的公式为:
T
1=(a
1,b
1,c
1)
其中,T
1为地面的法向量,α为翻滚角,β为俯仰角。
S610,根据激光雷达的安装高度、集卡托架高度、集装箱高度以及与激光雷达的距离,从第一方向标定三维点云中确定集装箱侧面点云。
集装箱侧面点云是指采集的集装箱作业现场的第一方向标定三维点云中表示集装箱侧面部分的点云。具体可根据点云高度,以及点云与激光雷达的距离确定。
具体地,集装箱侧面点云,已知激光雷达的高度为a,集卡托架高度为b,集装箱高度为c,取的z坐标范围为[-a+b,-a+b+c]的点云,作为一次过滤后的点云。由于集装箱侧面靠近激光雷达,设置距离阈值t,在一次过滤后的点云基础上,取与激光雷达的距离小于t的点云作为集装箱侧面点云。
S612,计算集装箱侧面点云的平面法向量。
集装箱侧面点云的平面法向量的计算方法与步骤S606相同,此处不再赘述。
S614,根据集装箱侧面点云的平面法向量,计算基准激光雷达的偏航角,第一姿态角包括翻滚角、俯仰角和偏航角。
其中,偏航角为激光雷达坐标系Z轴与集装箱侧面的夹角。
具体地,计算偏航角的计算公式为:
T
2=(a
2,b
2,c
2)
其中,T
2为集装箱侧面点云的平面法向量,γ为偏航角。
本实施例中,通过平面法向量的方法,计算激光雷达的翻滚角、俯仰角和偏航角。
在另一个实施例中,获取对准激光雷达的第二姿态角和对准激光雷达相对于基准激光雷达的位置平移量,包括:获取对同一标定物,基准激光雷达采集的标定物第一三维点云,以及对准激光雷达采集的标定物第二三维点云;将标定物第一三维点云转换至检测系统坐标系;对标定物第二三维点云和转换后的标定物第一三维点云进行点云匹配,确定对准激光雷达第二姿态角以及对准激光雷达相对于基准激光雷达的平移量。
其中,将标定物第一三维点云转换到检测系统坐标系,利用已标定的第一姿态角进行转换。具体地,通过根据转换后的基准激光雷达对对准激光雷达之间的姿态角与位置平移量进行标定,目的是将两个激光雷达标定,使其点云能转换到同一坐标系,减少检测系统的检测盲区。
通过在基准激光雷达与对准激光雷达的共视区域摆放具有特定形态的标定物,将基准激光雷达采集的标定物第一三维点云转换到检测系统坐标系,然后对转换后的标定物第一三维点云和对准激光雷达采集的标定物第二三维点云,利用点云匹配计算激光雷达间姿态,根据相同物体在不同坐标系下的点云数据差异,确定对准基光雷达的第二姿态角和对准激光雷达相对于基准激光雷达的平移量。其中,点云匹配的方法可使用的是常用的icp(Iterative Closest Point)、ndt(正态分布变换)等。
在步骤S406之后,还包括:S408,根据第一姿态角将第一三维点云转换至检测系统坐标系。
如前面的,以基准激光雷达为原点,建立检测系统坐标系。通过根据第一姿态角将第一三维点云转换至与检测系统坐标系的平面平行。
具体地,根据基准激光雷达的翻滚角和俯仰角,对第一三维点云进行转换,转换后的第一三维点云的地面点云与检测系统坐标系的XOY平面平行,根据基准激光雷达的偏航角,对转换后的第一三维点云进行转换,转换后的第一三维点云的集装箱侧面点云与检测系统坐标系的XOZ平行。
具体地,根据基准激光雷达的俯仰角,将第一三维点云绕检测系统坐标系的 X轴旋转,根据基准激光雷达的翻滚角,将第一三维点云绕检测系统坐标系的Y轴旋转,转换第一三维点云中的地面点云与激光雷达坐标系的底平面平行。如下所示:
p
g=R
y·R
x·p
c
其中,R
x和R
y为绕x轴与绕y轴的旋转矩阵,p
g为转换后与检测系统坐标系XOY平面平行的第一三维点云中的地面点云,p
c为原始地面点云。
具体地,根据基准激光雷达的偏航角,将转换后的第一三维点云绕检测系统坐标系的Z轴旋转,转换后第一三维点云中的集装箱侧面点云与激光雷达坐标系的侧平面平行。如下所示:
p=R
z·p
g
其中,R
z为绕z轴的旋转矩阵,p
g为转换后与地面与XOY平面平行的点云,p为最终转换后集装箱侧面点云与检测系统坐标系的XOZ平面平行的点云。
S410,根据第二姿态角和位置平移量将第二三维点云转换至检测系统坐标系。
具体地,根据第二姿态角和位置平移量确定对准激光雷达相对于已转换到检测系统坐标系的基准激光雷达的旋转矩阵,根据旋转矩阵将第二三维点云转换至与检测系统坐标系的平面平行。
其中,令α
l为翻滚角,β
l为俯仰角,γ
l为偏航角,T
l为位置平移量,则对准激光雷达相对于已转换到检测系统坐标系的基准激光雷达的旋转矩阵R为:
p
lg=R(α
l,β
l,
γ
l)·p
l
p
l为对准激光雷达采集的原始第二三维点云,p
lg为检测系统坐标系下转换后的第二三维点云。
S412,融合转换后的第一三维点云和转换后的第二三维点云,得到集装箱作业的全面点云。
基准激光雷达在检测系统坐标系下的第一三维点云为p
g,对准激光雷达在检测系统坐标系下的第二三维点云p
lg,则融合得到的全面点云p
R=p
g+p
lg。
S414,获取吊具与基准激光雷达的相对平移量。
具体地,吊具与基准激光雷达的相对平移量,反应了吊具与基准激光雷达的相对位置关系。本实施例中,手动测量吊具在缩紧状态,即位于最高位置时吊具最右下角到基准激光雷达的相对平移量T
d=(T
dx T
dy 0)。吊具与基准激光雷达的相对位置关系如图5所示,左侧为三维图,右侧为俯视图,此时吊具处于紧缩状态。如图6所示,D
x为吊具紧缩时的长,D
y为吊具的宽。
S416,根据相对平移量以及集装箱的尺寸参数,在全面点云中确定集装箱下落区域范围。
在实用应用中,集装箱以下的区域,都可以作为集装箱下落区域范围。
S418,当在集装箱下落区域范围内检测到障碍物时,发出防砸警报。
如前面所提及的,集装箱下落区域范围,是吊具夹住箱装箱下落时集装箱以下的区域。因此,可将在箱装箱下落区域进行障碍物检测,当在集装箱下落区域范围内检测到障碍物时,发出防砸警报。其中,障碍物是指集装箱下落区域范围内非集卡托架的一切物体。
一个实施例中,如图13所示,当在集装箱下落区域范围内检测到障碍物时,发出防砸警报的步骤包括:
S1102,根据集卡托架高度阈值,在全面点云中滤除集卡托架点云。
具体地,集卡托架高度区间波动幅度不大,可以根据经验值,设置集卡托架 高度阈值H
h,集卡托架点云即全面点云中Z坐标值小于高度阈值的点云,在全面点云p
R进行直通滤波去除集卡托架区域。集卡侧视图及其预警区域、高度阈值如图14所示。
S1104,对集装箱下落区域范围内的点云,进行滤波处理。
在接收到控制设备给出的吊具开始下落信号后,检测系统对集装箱下落区域范围内的点云p
X,进行点云去噪滤波处理。采用的去噪滤波算法为半径点滤波,根据空间点半径范围临近点数量来滤波,只有在一定范围内存在大于设定阈值数量的点云,才进行保留。
S1106,对滤波处理后集装箱下落区域范围内的点云,进行障碍物检测。
其中,对于滤波处理后的点云,可利用障碍物分类算法检测。
S1108,若检测到障碍物点云,则发出防砸警报。
若检测存在障碍物点云,则将判定有砸到的可能性,主控设备输出防砸警报信号给集装箱起重机的控制设备,进行防砸预警。
本申请的实施例中,通过滤除托架点云,在集装箱下落区域范围内进行障碍物检测。
在另一个实施例中,如图15所示,当在集装箱下落区域范围内检测到障碍物时,发出防砸警报,包括:
S1302,将全面点云投影为二维图像。
具体地,对于每一个全面点云,以像素点表示,得到二维图像。
如图16所示,将全面点云投影为二维图像的步骤包括:
S1402,对全面点云,计算各全面点云的二维坐标。
具体地,对于全面点云中的每一个三维点,可以以下公式计算其二维图像的坐标。
u=[(x
i-x
min)/u
r]
v=[(z
i-z
min)/v
r]
其中,u和v为二维图像的行坐标和列坐标,x
i和z
i为第i个全面点云的x轴坐标和z轴坐标,x
min和z
min为全面点云在x轴和Z轴的最小值,u
r和v
r为全面点云投影到二维图像上的精度,代表了二维图像上相邻像素点之间的实际距离。
S1404,根据各全面点云的二维坐标,将点云转换为像素点。
具体地,以像素点表示全面点云,像素点的坐标即为全面点云的二维坐标。
S1406,将点云像素点和非点云像素点进行二值化处理,得到二值图像。
具体地,二值化处理是指将图像上的像素点的灰度值设置为0或255,也就是将整个图像呈现出明显的黑白效果的过程。一种方式可以为,将从点云转换的像素点的灰度值设为255,其它的非点云转换的像素点的灰度值设为0,得到二值图像。另一种方式可以为,将从点云转换的像素点的灰度值设为0,其它的非点云转换的像素点的灰度值设为255,得到二值图像。
S1408,对二值图像进行图像预处理,得到二维图像。
其中,图像的预处理包括:首先对二维图像进行中值滤波和双边滤波预处理操作,中值滤波是为了保护边缘信息,双边滤波是为了保边去噪;然后进行形态学膨胀操作。由于激光传感器的扫描方式,有些临近点之间的距离会大于图像的像素距离,导致图像出现孔洞,如果增大像素精度,又会降低图像的分辨率,在图像上进行膨胀操作能有效的减少孔洞。
图像预处理方法,不仅限于形态学膨胀。也可以对图像进行形态学闭运算,以填充黑洞区域,然后进行形态学开运算,以增强边缘信息,过滤离散的干扰像素点。
在步骤S1302之后,还包括:
S1304,确定集装箱下落区域范围以及集卡托架在二维图像中的位置范围。
具体地,利用上述的三维点云的二维图像的坐标的公式,对集装箱下落区域范围的三维点云,计算其对应的二维坐标,得到集装箱下落区域范围在二维图像中的位置范围。根据集卡托架高度阈值,利用上述的三维点云的二维图像的坐标的公式,对集卡托架的三维点云,计算其对应的二维坐标,得到集卡托架在二维图像中的位置范围。如图17代表了集装箱下方无障碍物的二维图像示意图,图18代表了集装箱下方有障碍物的二维图像示意图。
S1306,根据确定的集卡托架在二维图像中的位置范围,去除二维图像中的集卡托架像素点。
S1308,对二维图像中的集装箱下落区域范围进行图像检测,若检测到障碍 物,发出防砸警报。
具体地,对二维图像中的集装箱下落区域范围的像素点进行图像检测,得到集卡防砸检测结果。
一种实施方式,如图19所示,对二维图像中的集装箱下落区域的位置范围的像素点进行图像检测,若检测到障碍物,发出防砸警报的步骤,包括:
S1702,遍历二维图像中集装箱下落区域的位置范围内的各行,统计各行中点云像素点的数量。
其中,点云像素点是指从点云转换的像素点。根据二值化的规则,点云像素点的灰度值可以为255,非点云像素点的灰度值为0。点云像素点的灰度值可以为0,非点云像素点的灰度值为255。具体地,根据点云像素点的灰度值,统计二维图像中集装箱下落区域的位置范围内的各行中像素点的灰度值为相应数值的像素点数量。例如,点云像素点的灰度值为255,则统计二维图像的各行中像素点的灰度值为255的像素点数量,即统计每行中有多少个像素点的像素值为255,从而得到每行中点云像素点的数量。
S1704,将当前行点云像素点的数量与第一阈值进行比较。
若当前行点云像素点的数量大于第一阈值,则执行步骤S1706,若当前行点云像素点的数量小于第一阈值,则执行步骤S1708。
S1706,计数器增加预设值。
具体地,预设值为1,若当前行点云像素点的数量大于第一阈值,则计数器加1。步骤S1706之后执行步骤S1708。
S1708,判断集装箱下落区域的位置范围内的各行是否遍历完成。
若是,则执行步骤S1710,若否,则返回步骤S1702。
S1710,将计数器的统计值与第二阈值进行比较。
若计数器的统计值大于第二阈值,则执行步骤S1712。
S1712,得到检测到障碍物的检测结果,发出防砸警报。
其中,如图17所示,集装箱下落区域无障碍物,集装箱下落区域的位置范围当未采集到三维点云,对应地,集装箱下落区域内的各行中点云像素点的数量为0。如图18所示,集装箱下落区域有障碍物,集装箱下落区域的位置范围采 集到三维点云,如图头集卡车头部分位于集装箱下落区域的位置范围内,集装箱下落区域的位置范围内有部分行的像素个数都大于0,并且超出阈值T1的行数,即计数器的统计值大于T2,则可以检测到障碍物。第一阈值和第二阈值可以根据精度要求和经验值设定。
该方法可应用在集装箱码头的集装箱门吊设备上,吊具吊着集装箱下落装载集卡时,系统能判断集卡车头或托架上其他集装箱被下落集装箱砸到的可能性,避免发生事故,适应集卡带20ft、双20ft、40ft和45ft集装箱状态。
应该理解的是,虽然上述各流程图中的各个步骤按照箭头的指示依次显示,但是这些步骤并不是必然按照箭头指示的顺序依次执行。
在一个实施例中,如图20所示,提供了一种基于三维激光的集卡防砸检测装置,装置包括:集装箱获取模块2002,用于获取集卡装箱作业中集装箱起重机的吊具当前所夹住的集装箱的尺寸参数;点云获取模块2004,获取在吊具夹住集装箱下落时激光雷达采集的集装箱作业的三维点云;姿态参数获取模块2006,用于获取激光雷达的姿态参数。转换模块2008,用于根据姿态参数,对三维点云进行转换,得到集装箱作业的全面点云;位置获取模块2010,用于获取吊具与激光雷达的相对平移量;下落区域确定模块2012,用于根据相对平移量以及集装箱的尺寸参数,在全面点云中确定集装箱下落区域范围;检测模块2014,用于当在集装箱下落区域范围内检测到障碍物时,发出防砸警报。
在另一个实施例中,激光雷达包括安装在集装箱起重机吊具下方的同一侧呈一定夹角设置的基准激光雷达和对准激光雷达;姿态参数包括基准激光雷达的第一姿态角,对准激光雷达的第二姿态角和对准激光雷达相对于基准激光雷达的位置平移量。
点云获取模块,用于获取在吊具夹住集装箱下落时基准激光雷达采集的集装箱作业的第一三维点云,以及对准激光雷达采集的集装箱作业的第二三维点云。
转换模块,包括:第一转换模块,用于根据第一姿态角将第一三维点云转换至检测系统坐标系;第二转换模块,用于根据第二姿态角和位置平移量将第二三维点云转换至检测系统坐标系;融合模块,用于融合转换后的第一三维点云和转 换后的第二三维点云,得到集装箱作业的全面点云;位置获取模块,用于获取吊具与基准激光雷达的相对平移量。
在另一个实施例中,姿态参数获取模块,包括:标定点云获取模块,用于获取在标定状态吊具夹住集装箱下落时基准激光雷达采集的第一方向标定三维点云。地面点云确定模块,用于根据基准激光雷达的安装高度,从第一方向标定三维点云中确定地面点云;法向量计算模块,用于计算地面点云的平面法向量;角度确定模块,用于根据地面点云的平面法向量,计算基准激光雷达的翻滚角和俯仰角;侧面点云确定模块,用于根据基准激光雷达的安装高度、集卡托架高度、集装箱高度以及与基准激光雷达的距离,从第一方向标定三维点云中确定集装箱侧面点云;法向量计算模块,还用于计算集装箱侧面点云的平面法向量;角度确定模块,还用于根据集装箱侧面点云的平面法向量,计算基准激光雷达的偏航角,第一姿态角包括翻滚角、俯仰角和偏航角。
在另一个实施例中,姿态参数获取模块还包括:标定模块,用于获取对同一标定物,对准激光雷达采集的标定物第一三维点云,以及基准激光雷达采集的标定物第二三维点云;第一转换模块,还用于将标定物第一三维点云转换至检测系统坐标系;匹配模块,用于对转换后的标定物第一三维点云和标定物第二三维点云进行点云匹配,确定对准激光雷达的第二姿态角以及位置平移量。
在另一个实施例中,检测模块,包括:点云滤除模块,用于根据集卡托架高度阈值,在全面点云中滤除集卡托架点云;滤波处理模块,用于对集装箱下落区域范围内的点云,进行滤波处理;障碍物检测模块,用于对滤波处理后集装箱下落区域范围内的点云,进行障碍物检测;警报模块,用于若检测到障碍物点云,则发出防砸警报。
在另一个实施例中,检测模块,还包括:投影模块,用于将全面点云投影为二维图像;位置确定模块,用于确定集装箱下落区域范围以及集卡托架在二维图像中的位置范围;像素点滤除模块,用于根据确定的集卡托架在二维图像中的位置范围,去除二维图像中的集卡托架像素点;像素点检测模块,用于对二维图像中的集装箱下落区域范围的像素点进行图像检测,若检测到障碍物,发出防砸警报。
在另一个实施例中,投影模块,包括:坐标计算模块,用于对全面点云,计算各全面点云的二维坐标;像素点转换模块,用于根据各全面点云的二维坐标,将点云转换为像素点;二值化模块,用于将点云像素点和非点云像素点进行二值化处理,得到二值图像;预处理模块,用于对二值图像进行图像预处理,得到二维图像。
在另一个实施例中,像素点检测模块包括:遍历模块,用于遍历二维图像中集装箱下落区域的位置范围内的各行,统计各行中点云像素点的数量;计数器,用于若当前行点云像素点的数量大于第一阈值,则计数器增加预设值;比较模块,用于在集装箱下落区域的位置范围内的各行遍历完成后,将计数器的统计值与第二阈值进行比较;防砸检测模块,用于若计数器的统计值大于第二阈值,则得到检测到障碍物的检测结果,发出防砸警报。
关于基于三维激光的集卡防砸检测装置的具体限定可以参见上文中对于基于三维激光的集卡防砸检测方法的限定,在此不再赘述。上述基于三维激光的集卡防砸检测装置中的各个模块可全部或部分通过软件、硬件及其组合来实现。上述各模块可以硬件形式内嵌于或独立于计算机设备中的处理器中,也可以以软件形式存储于计算机设备中的存储器中,以便于处理器调用执行以上各个模块对应的操作。
在一个实施例中,提供了一种计算机设备,该计算机设备可以是图1中的主控设备,其内部结构图可以如图21所示。该计算机设备包括通过系统总线连接的处理器、存储器、通信接口、显示屏和输入装置。其中,该计算机设备的处理器用于提供计算和控制能力。该计算机设备的存储器包括非易失性存储介质、内存储器。该非易失性存储介质存储有操作系统和计算机程序。该内存储器为非易失性存储介质中的操作系统和计算机程序的运行提供环境。该计算机设备的通信接口用于与外部的终端进行有线或无线方式的通信,无线方式可通过WIFI、运营商网络、NFC(近场通信)或其他技术实现。该计算机程序被处理器执行时以实现一种基于三维激光的集卡防砸检测方法。该计算机设备的显示屏可以是液晶显示屏或者电子墨水显示屏,该计算机设备的输入装置可以是显示屏上覆盖的触摸层,也可以是计算机设备外壳上设置的按键、轨迹球或触控板,还可以 是外接的键盘、触控板或鼠标等。
本领域技术人员可以理解,图21中示出的结构,仅仅是与本申请方案相关的部分结构的框图,并不构成对本申请方案所应用于其上的计算机设备的限定。
在一个实施例中,提供了一种计算机设备,包括存储器和处理器,存储器中存储有计算机程序,该处理器执行计算机程序时实现上述各实施例的三维激光的集卡防砸检测方法的步骤。
在一个实施例中,提供了一种计算机可读存储介质,其上存储有计算机程序,计算机程序被处理器执行时实现以下上述各实施例的三维激光的集卡防砸检测方法的步骤。
本领域普通技术人员可以理解实现上述实施例方法中的全部或部分流程,是可以通过计算机程序来指令相关的硬件来完成,的计算机程序可存储于一非易失性计算机可读取存储介质中,该计算机程序在执行时,可包括如上述各方法的实施例的流程。其中,本申请所提供的各实施例中所使用的对存储器、存储、数据库或其它介质的任何引用,均可包括非易失性和易失性存储器中的至少一种。非易失性存储器可包括只读存储器(Read-Only Memory,ROM)、磁带、软盘、闪存或光存储器等。易失性存储器可包括随机存取存储器(Random Access Memory,RAM)或外部高速缓冲存储器。作为说明而非局限,RAM可以是多种形式,比如静态随机存取存储器(Static Random Access Memory,SRAM)或动态随机存取存储器(Dynamic Random Access Memory,DRAM)等。
以上实施例的各技术特征可以进行任意的组合,为使描述简洁,未对上述实施例中的各个技术特征所有可能的组合都进行描述,然而,只要这些技术特征的组合不存在矛盾,都应当认为是本说明书记载的范围。
以上实施例仅表达了本申请的几种实施方式,其描述较为具体和详细,但并不能因此而理解为对发明专利范围的限制。应当指出的是,对于本领域的普通技术人员来说,在不脱离本申请构思的前提下,还可以做出若干变形和改进,这些都属于本申请的保护范围。因此,本申请专利的保护范围应以所附权利要求为准。
Claims (11)
- 一种基于三维激光的集卡防砸检测方法,所述方法包括:获取集卡装箱作业中集装箱起重机的吊具当前所夹住的集装箱的尺寸参数;获取在所述吊具夹住所述集装箱下落时激光雷达采集的集装箱作业的三维点云;获取所述激光雷达的姿态参数;根据所述姿态参数,对所述三维点云进行转换,得到集装箱作业的全面点云;获取吊具与所述激光雷达的相对平移量;根据所述相对平移量以及所述集装箱的尺寸参数,在所述全面点云中确定集装箱下落区域范围;当在所述集装箱下落区域范围内检测到障碍物时,发出防砸警报。
- 根据权利要求1所述的方法,其特征在于,所述激光雷达包括安装在集装箱起重机吊具下方的同一侧呈一定夹角设置的基准激光雷达和至少一个对准激光雷达;所述姿态参数包括所述基准激光雷达的第一姿态角、所述对准激光雷达的第二姿态角和所述对准激光雷达相对于所述基准激光雷达的位置平移量;获取在所述吊具夹住所述集装箱下落时激光雷达采集的集装箱作业的三维点云,包括:获取在所述吊具夹住所述集装箱下落时所述基准激光雷达采集的集装箱作业的第一三维点云,以及对准激光雷达采集的集装箱作业的第二三维点云;根据所述姿态参数,对所述三维点云进行转换,得到集装箱作业的全面点云,包括:根据所述第一姿态角将所述第一三维点云转换至检测系统坐标系;根据所述第二姿态角和所述位置平移量将所述第二三维点云转换至所述检测系统坐标系;融合转换后的所述第一三维点云和转换后的所述第二三维点云,得到集装箱作业的全面点云;其中,所述相对平移量为所述吊具与所述基准激光雷达的相对平移量。
- 根据权利要求2所述的方法,其特征在于,获取所述基准激光雷达的第一姿态角的方式,包括:获取在标定状态所述吊具夹住所述集装箱下落时所述基准激光雷达采集的第一方向标定三维点云;根据所述基准激光雷达的安装高度,从所述第一方向标定三维点云中确定地面点云;计算所述地面点云的平面法向量;根据所述地面点云的平面法向量,计算所述基准激光雷达的翻滚角和俯仰角;根据所述基准激光雷达的安装高度、集卡托架高度、集装箱高度以及与所述基准激光雷达的距离,从所述第一方向标定三维点云中确定集装箱侧面点云;计算所述集装箱侧面点云的平面法向量;根据所述集装箱侧面点云的平面法向量,计算所述基准激光雷达的偏航角,所述第一姿态角包括所述翻滚角、所述俯仰角和所述偏航角。
- 根据权利要求2所述的方法,其特征在于,获取对准激光雷达的第二姿态角和所述对准激光雷达相对于所述基准激光雷达的位置平移量的方式,包括:获取对同一标定物,所述基准激光雷达采集的标定物第一三维点云,以及所述对准激光雷达采集的标定物第二三维点云;将所述标定物第一三维点云转换至检测系统坐标系;对所述标定物第二三维点云和转换后的所述标定物第一三维点云进行点云匹配,确定所述对准激光雷达的第二姿态角以及所述对准激光雷达相对于所述基准激光雷达的位置平移量。
- 根据权利要求1所述的方法,其特征在于,当在所述集装箱下落区域范围内检测到障碍物时,发出防砸警报,包括:根据集卡托架高度阈值,在所述全面点云中滤除集卡托架点云;对所述集装箱下落区域范围内的点云,进行滤波处理;对滤波处理后所述集装箱下落区域范围内的点云,进行障碍物检测;若检测到障碍物点云,则发出防砸警报。
- 根据权利要求1所述的方法,其特征在于,当在所述集装箱下落区域范围内检测到障碍物时,发出防砸警报,包括:将所述全面点云投影为二维图像;确定所述集装箱下落区域范围以及集卡托架在所述二维图像中的位置范围;根据确定的集卡托架在二维图像中的位置范围,去除二维图像中的集卡托架像素点;对所述二维图像中的所述集装箱下落区域范围的像素点进行图像检测,若检测到障碍物,发出防砸警报。
- 根据权利要求6所述的方法,其特征在于,所述将所述全面点云投影为二维图像,包括:对所述全面点云,计算各全面点云的二维坐标;根据各全面点云的二维坐标,将点云转换为像素点;将点云像素点和非点云像素点进行二值化处理,得到二值图像;对所述二值图像进行图像预处理,得到二维图像。
- 根据权利要求7所述的方法,其特征在于,对所述二维图像中的所述集装箱下落区域的位置范围的像素点进行图像检测,若检测到障碍物,发出防砸警报,包括:遍历所述二维图像中所述集装箱下落区域的位置范围内的各行,统计各行中点云像素点的数量;若当前行点云像素点的数量大于第一阈值,则计数器增加预设值;在所述集装箱下落区域的位置范围内的各行遍历完成后,将计数器的统计值与第二阈值进行比较;若所述计数器的统计值大于所述第二阈值,则得到检测到障碍物的检测结果,发出防砸警报。
- 一种基于三维激光的集卡防砸检测装置,其特征在于,所述装置包括:集装箱获取模块,用于获取集卡装箱作业中集装箱起重机的吊具当前所夹住的集装箱的尺寸参数;点云获取模块,获取在所述吊具夹住所述集装箱下落时激光雷达采集的集装箱作业的三维点云;姿态参数获取模块,用于获取所述激光雷达的姿态参数;转换模块,用于根据所述姿态参数,对所述三维点云进行转换,得到集装箱作业的全面点云;位置获取模块,用于获取吊具与所述基准激光雷达的相对平移量;下落区域确定模块,用于根据所述相对平移量以及所述集装箱的尺寸参数,在所述全面点云中确定集装箱下落区域范围;检测模块,用于当在所述集装箱下落区域范围内检测到障碍物时,发出防砸警报。
- 一种计算机设备,包括存储器和处理器,所述存储器存储有计算机程序,其特征在于,所述处理器执行所述计算机程序时实现权利要求1至8中任一项所述方法的步骤。
- 一种计算机可读存储介质,其上存储有计算机程序,其特征在于,所述计算机程序被处理器执行时实现权利要求1至8中任一项所述的方法的步骤。
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010158561.XA CN113376654B (zh) | 2020-03-09 | 2020-03-09 | 基于三维激光的集卡防砸检测方法、装置和计算机设备 |
CN202010158561.X | 2020-03-09 |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2021179988A1 true WO2021179988A1 (zh) | 2021-09-16 |
Family
ID=77568530
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/CN2021/079102 WO2021179988A1 (zh) | 2020-03-09 | 2021-03-04 | 基于三维激光的集卡防砸检测方法、装置和计算机设备 |
Country Status (2)
Country | Link |
---|---|
CN (1) | CN113376654B (zh) |
WO (1) | WO2021179988A1 (zh) |
Cited By (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113814953A (zh) * | 2021-10-13 | 2021-12-21 | 国网山西省电力公司输电检修分公司 | 一种巡检机器人的履带自动找平方法、系统及电子设备 |
CN114782438A (zh) * | 2022-06-20 | 2022-07-22 | 深圳市信润富联数字科技有限公司 | 物体点云修正方法、装置、电子设备和存储介质 |
CN115797359A (zh) * | 2023-02-10 | 2023-03-14 | 苏州赫芯科技有限公司 | 基于电路板上锡膏的检测方法、设备和存储介质 |
CN115849189A (zh) * | 2022-11-16 | 2023-03-28 | 上海西井信息科技有限公司 | 基于点云的吊具二次锚定方法、系统、设备及存储介质 |
CN115893032A (zh) * | 2022-11-23 | 2023-04-04 | 哈尔滨工业大学芜湖机器人产业技术研究院 | 用于厢式集装箱的自动装车系统 |
WO2023070954A1 (zh) * | 2021-10-29 | 2023-05-04 | 北京航天自动控制研究所 | 一种基于机器视觉的集卡引导和单双箱识别方法和装置 |
CN117908045A (zh) * | 2024-01-22 | 2024-04-19 | 深圳市平方科技股份有限公司 | 一种船上人员防跌落检测方法及相关设备 |
Families Citing this family (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN114701986B (zh) * | 2022-03-10 | 2024-03-22 | 杭州海康机器人股份有限公司 | 执行机构的控制方法和控制装置 |
CN115077385B (zh) * | 2022-07-05 | 2023-09-26 | 北京斯年智驾科技有限公司 | 无人集卡集装箱位姿测量方法及其测量系统 |
CN115930791B (zh) * | 2022-12-02 | 2024-01-12 | 湖北润铁轨道装备有限公司 | 一种多模态数据的集装箱货物位置与尺寸检测方法 |
CN115641553B (zh) * | 2022-12-26 | 2023-03-10 | 太原理工大学 | 掘进机作业环境侵入物的在线检测装置及方法 |
Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2006160402A (ja) * | 2004-12-03 | 2006-06-22 | Mitsui Eng & Shipbuild Co Ltd | コンテナクレーンにおけるシャシ位置検出装置 |
CN104085794A (zh) * | 2014-07-14 | 2014-10-08 | 上海海事大学 | 一种用于集装箱分布轮廓和位置智能化检测系统及方法 |
US20150191333A1 (en) * | 2011-05-10 | 2015-07-09 | Cargotec Finland Oy | System for determination of a container's position in a vehicle and/or its trailer to be loaded with containers |
CN106794971A (zh) * | 2014-06-30 | 2017-05-31 | 科尼全球公司 | 通过负载处理装备的负载运输 |
CN107089599A (zh) * | 2017-06-29 | 2017-08-25 | 北京国泰星云科技有限公司 | 基于集装箱箱区轮廓三维识别的吊具安全防撞系统及方法 |
CN109752726A (zh) * | 2019-01-23 | 2019-05-14 | 上海海事大学 | 一种集装箱姿态检测装置和方法 |
CN110431101A (zh) * | 2017-03-16 | 2019-11-08 | 科尼起重机全球公司 | 将集装箱下降到运输平台或提升集装箱远离运输平台时集装箱转移装置的监控 |
CN110673183A (zh) * | 2019-09-24 | 2020-01-10 | 南通润邦重机有限公司 | 一种结合gps\ins的集装箱识别定位方法 |
-
2020
- 2020-03-09 CN CN202010158561.XA patent/CN113376654B/zh active Active
-
2021
- 2021-03-04 WO PCT/CN2021/079102 patent/WO2021179988A1/zh active Application Filing
Patent Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2006160402A (ja) * | 2004-12-03 | 2006-06-22 | Mitsui Eng & Shipbuild Co Ltd | コンテナクレーンにおけるシャシ位置検出装置 |
US20150191333A1 (en) * | 2011-05-10 | 2015-07-09 | Cargotec Finland Oy | System for determination of a container's position in a vehicle and/or its trailer to be loaded with containers |
CN106794971A (zh) * | 2014-06-30 | 2017-05-31 | 科尼全球公司 | 通过负载处理装备的负载运输 |
CN104085794A (zh) * | 2014-07-14 | 2014-10-08 | 上海海事大学 | 一种用于集装箱分布轮廓和位置智能化检测系统及方法 |
CN110431101A (zh) * | 2017-03-16 | 2019-11-08 | 科尼起重机全球公司 | 将集装箱下降到运输平台或提升集装箱远离运输平台时集装箱转移装置的监控 |
CN107089599A (zh) * | 2017-06-29 | 2017-08-25 | 北京国泰星云科技有限公司 | 基于集装箱箱区轮廓三维识别的吊具安全防撞系统及方法 |
CN109752726A (zh) * | 2019-01-23 | 2019-05-14 | 上海海事大学 | 一种集装箱姿态检测装置和方法 |
CN110673183A (zh) * | 2019-09-24 | 2020-01-10 | 南通润邦重机有限公司 | 一种结合gps\ins的集装箱识别定位方法 |
Non-Patent Citations (1)
Title |
---|
HUNAGHE CAI, SHUANGHUA YANG, ZHENSHENG YU: "Locatization System of Automatic Rail-mounted Gantry Crane at Container Terminal", 31 December 2017 (2017-12-31), pages 21 - 24, XP055846579, ISSN: 1000-8969 * |
Cited By (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113814953A (zh) * | 2021-10-13 | 2021-12-21 | 国网山西省电力公司输电检修分公司 | 一种巡检机器人的履带自动找平方法、系统及电子设备 |
CN113814953B (zh) * | 2021-10-13 | 2024-03-19 | 国网山西省电力公司超高压变电分公司 | 一种巡检机器人的履带自动找平方法、系统及电子设备 |
WO2023070954A1 (zh) * | 2021-10-29 | 2023-05-04 | 北京航天自动控制研究所 | 一种基于机器视觉的集卡引导和单双箱识别方法和装置 |
CN114782438A (zh) * | 2022-06-20 | 2022-07-22 | 深圳市信润富联数字科技有限公司 | 物体点云修正方法、装置、电子设备和存储介质 |
CN114782438B (zh) * | 2022-06-20 | 2022-09-16 | 深圳市信润富联数字科技有限公司 | 物体点云修正方法、装置、电子设备和存储介质 |
CN115849189A (zh) * | 2022-11-16 | 2023-03-28 | 上海西井信息科技有限公司 | 基于点云的吊具二次锚定方法、系统、设备及存储介质 |
CN115849189B (zh) * | 2022-11-16 | 2024-01-30 | 上海西井科技股份有限公司 | 基于点云的吊具二次锚定方法、系统、设备及存储介质 |
CN115893032A (zh) * | 2022-11-23 | 2023-04-04 | 哈尔滨工业大学芜湖机器人产业技术研究院 | 用于厢式集装箱的自动装车系统 |
CN115797359A (zh) * | 2023-02-10 | 2023-03-14 | 苏州赫芯科技有限公司 | 基于电路板上锡膏的检测方法、设备和存储介质 |
CN117908045A (zh) * | 2024-01-22 | 2024-04-19 | 深圳市平方科技股份有限公司 | 一种船上人员防跌落检测方法及相关设备 |
Also Published As
Publication number | Publication date |
---|---|
CN113376654A (zh) | 2021-09-10 |
CN113376654B (zh) | 2023-05-26 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
WO2021179988A1 (zh) | 基于三维激光的集卡防砸检测方法、装置和计算机设备 | |
US20220270293A1 (en) | Calibration for sensor | |
WO2021179983A1 (zh) | 基于三维激光的集卡防吊起检测方法、装置和计算机设备 | |
CN113657224B (zh) | 车路协同中用于确定对象状态的方法、装置、设备 | |
WO2021197345A1 (zh) | 一种基于激光雷达的封闭空间内剩余体积的检测方法和装置 | |
CN111508027B (zh) | 摄像机外参标定的方法和装置 | |
CN110826499A (zh) | 物体空间参数检测方法、装置、电子设备及存储介质 | |
US20140203959A1 (en) | Object recognition system having radar and camera input | |
US20140205139A1 (en) | Object recognition system implementing image data transformation | |
CN105157608A (zh) | 一种超限车辆的检测方法、装置及系统 | |
JPWO2016199366A1 (ja) | 寸法測定装置および寸法測定方法 | |
WO2022183685A1 (zh) | 目标检测方法、电子介质和计算机存储介质 | |
CN114494075A (zh) | 基于三维点云的障碍物识别方法、电子设备和存储介质 | |
CN112799091A (zh) | 算法评估方法、装置及存储介质 | |
US20190195617A1 (en) | Container auto-dimensioning | |
US20230162442A1 (en) | Image processing apparatus, image processing method, and storage medium | |
CN114742726A (zh) | 盲区的检测方法、装置、电子设备及存储介质 | |
CN114493967A (zh) | 图像获取装置及方法、图像处理装置及方法、图像处理系统 | |
JP6561688B2 (ja) | 検出装置、検出方法、撮像装置、機器制御システム、及びプログラム | |
CN115077385B (zh) | 无人集卡集装箱位姿测量方法及其测量系统 | |
CN113643358B (zh) | 相机的外参标定方法、装置、存储介质及系统 | |
EP4216177A1 (en) | Image processing device of person detection system | |
CN116263952A (zh) | 车斗测量方法、设备、系统及存储介质 | |
JPH0681182U (ja) | 車両周辺監視装置 | |
CN118351500A (zh) | 智能路锥识别行驶路线的方法及装置 |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 21768259 Country of ref document: EP Kind code of ref document: A1 |
|
NENP | Non-entry into the national phase |
Ref country code: DE |
|
122 | Ep: pct application non-entry in european phase |
Ref document number: 21768259 Country of ref document: EP Kind code of ref document: A1 |