CN111652936B - Three-dimensional sensing and stacking planning method and system for open container loading - Google Patents
Three-dimensional sensing and stacking planning method and system for open container loading Download PDFInfo
- Publication number
- CN111652936B CN111652936B CN202010780533.1A CN202010780533A CN111652936B CN 111652936 B CN111652936 B CN 111652936B CN 202010780533 A CN202010780533 A CN 202010780533A CN 111652936 B CN111652936 B CN 111652936B
- Authority
- CN
- China
- Prior art keywords
- container
- rows
- stacking
- bags
- background
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/70—Determining position or orientation of objects or cameras
- G06T7/73—Determining position or orientation of objects or cameras using feature-based methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F7/00—Methods or arrangements for processing data by operating upon the order or content of the data handled
- G06F7/76—Arrangements for rearranging, permuting or selecting data according to predetermined rules, independently of the content of the data
- G06F7/78—Arrangements for rearranging, permuting or selecting data according to predetermined rules, independently of the content of the data for changing the order of data flow, e.g. matrix transposition or LIFO buffers; Overflow or underflow handling therefor
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/187—Segmentation; Edge detection involving region growing; involving region merging; involving connected component labelling
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/194—Segmentation; Edge detection involving foreground-background segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/60—Analysis of geometric attributes
- G06T7/62—Analysis of geometric attributes of area, perimeter, diameter or volume
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10028—Range image; Depth image; 3D point clouds
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Geometry (AREA)
- Mathematical Physics (AREA)
- General Engineering & Computer Science (AREA)
- Length Measuring Devices By Optical Means (AREA)
Abstract
The invention provides a three-dimensional sensing and stacking planning method and a system for loading of an open container. The three-dimensional sensing device has the characteristics of high precision and high speed, and the stacking planning supports various stacking modes such as pattern stacking and the like, so that the automation level and efficiency of loading the open container are greatly improved.
Description
Technical Field
The invention belongs to the field of machine vision perception, and further relates to a three-dimensional perception and stacking planning method and system for loading of an open container.
Background
An open container truck is a widely used transport vehicle that many items of daily life need to be transported, such as white granulated sugar, fertilizers, cement, flour, etc. The weight of a single bag of the object is large generally, the transportation amount of single loading is also large, the workload of workers is large and the loading efficiency is low in the traditional manual loading method, and in addition, the severe operation environment with high temperature, high humidity and high dust raising gradually appears in recent years, so that a manufacturer pays high loading cost every year.
With the progress of the automation technology in China, more and more automatic loading equipment is available, and the automatic loading of the open container truck is completed by a mechanical method. The equipment meets the requirement of automatic loading of single-bag heavy objects to a certain extent, and improves the quality and efficiency of loading. However, the existing automatic loading equipment lacks three-dimensional perception of open container trucks, or a truck driver needs to accurately park the trucks at a specific area position, or the posture of the automatic loading equipment needs to be manually adjusted according to the parking postures of the trucks, so that the automatic loading equipment is not high in automation level, cannot be well adapted to the types of the trucks, and is not particularly convenient to use.
Aiming at the problems and the requirements, the patent provides a three-dimensional perception of the position, the posture and the size of an open type container and a bagged object stacking planning method. The method can lead the automatic loading equipment to automatically sense different poses and different types of trucks parked by the trucks, automatically plan a stacking loading method according to the loading quantity, and visually servo-control the loading equipment to finish automatic loading. The whole loading process does not need manual intervention completely, the automation level and efficiency of loading the open type container are greatly improved, and the cost is saved for manufacturers.
Disclosure of Invention
Aiming at the problems and requirements, the invention provides a three-dimensional sensing method for the position, the posture and the size of an open type container and a bagged object stacking planning method, and the automation level and the efficiency of loading the open type container are greatly improved.
In order to achieve the purpose, the technical scheme of the invention is realized as follows:
a three-dimensional sensing and stacking planning method for loading of an open container includes the steps that a 2D laser scanning radar is carried on a transverse moving synchronous belt and moves along with the synchronous belt to serve as a sensing unit to collect three-dimensional laser point cloud data, and the method further includes the following steps:
s1, automatically calibrating to obtain a coordinate conversion relation between a sensing unit coordinate system and a loading unit coordinate system;
s2, background modeling, filtering based on the background model, and filtering background points;
s3, preprocessing data, and performing noise reduction processing on the point cloud data obtained in the step S2;
s4, carrying out container size detection and pose estimation according to the preprocessed point cloud data;
and S5, carrying out stacking planning according to the size and the pose of the cargo box obtained in the S4 and the size information of the known cargo.
Further, the method for automatic calibration in step S1 includes:
s11, mounting an artificial marker on a fixed object in the motion range of the loading unit;
s12, controlling the loading unit to move to the position of the artificial marker in the loading operation space through an instruction, and respectively acquiring a coordinate P _ sense of the artificial marker in a sensing unit coordinate system and a coordinate P _ load in a loading unit coordinate system;
s13, constructing a homogeneous coordinate matrix P _ S _ matrix of P _ sense; and (3) constructing a coordinate matrix P _ l _ matrix of the P _ load, calculating a generalized inverse matrix P _ s _ matrix _ inv of the P _ s _ matrix, wherein a conversion matrix of a sensing unit coordinate system and a loading unit coordinate system is [ R t ] = P _ l _ matrix x P _ s _ matrix _ inv, calculating an optimal coordinate system conversion matrix by using a least square method, and completing automatic calibration.
Further, in step S2, the method specifically includes:
s21, background offline modeling: acquiring three-dimensional laser point cloud data of a background, calculating a vertical normal vector and curvature of each point of the point cloud, dividing the background point cloud data into a plurality of surface elements by using a region growing algorithm, wherein each surface element has similar vertical normal and curvature, calculating a central point, a normal direction and a maximum radius of each surface element as characteristics, and sequentially storing the characteristics in a background model configuration file to complete background modeling;
s22, background online filtering: loading a background model; after point cloud data containing the open container truck is obtained, the distance from each point in the current point cloud to each background surface element is calculated to judge whether the point is in the size range of the background surface element, then whether the current point belongs to the background is judged according to a specific threshold value, if the current point belongs to the background, the point is filtered, and the process is repeated until all the points complete background filtering.
Further, in step S4, the specific method for container detection and pose estimation includes:
s41, segmenting the point cloud of the container by using a region growing algorithm to obtain each surface element of the open container;
s42, calculating the geometric dimension information of each surface element by using a minimum bounding box algorithm;
and S43, determining coordinates of each vertex of the container in the coordinate system of the sensing unit according to the size of the open container, and calculating the pose of the open container.
Further, the specific method for pallet planning in step S5 includes:
a. calculating the total bag number of the loading according to the total weight of the loading and the weight of each bag of the bagged powdery materials: n = Weight/Weight;
b. determining the overlapping distance of the article bags and the distance extended by which the side edge of the article bag can extend out of the carriage;
c. the stacking plan is divided into a container inner part, a container upper part and a top layer part from bottom to top, and article bags in the container inner part cannot be stretched out, so that the extended = 0; the top layer part is used for ensuring stability and needs to be stacked on a seam;
d. the stacking is divided into three modules: in the container, the related parameters are expressed by in; an upper part of the cargo box, and relevant parameters are expressed by out; the top layer of the container, and the related parameters are represented by top; rows represent rows, cols represent columns, and layers represent palletizing layer numbers;
n represents the total number of bags needing loading, and N _ in represents the number of bags which can be stacked in the container;
the stacking planning adopts integer linear planning, firstly, part of a container is planned, the number of goods bags which can be stacked in each row in the container is calculated, namely cols _ in = round (L/(w-overlap/2)), L is the length of a carriage, w is the width of the goods bag, and round (#) is a rounding function;
calculating the number rows _ in = round (W/(l-overlap)), where each row in the container can be stacked, W is the width of the carriage, and l is the length of the article bag;
calculating the number of layers which can be stacked in the container, wherein layers _ in = round (H/H), H is the height of the compartment, and H is the height of the goods bag;
calculating the total number of bags which can be stacked in the container, wherein N _ in = layers _ in _ cols _ in _ rows _ in;
e. if N _ in < N, the goods are not completely loaded, and the goods are continuously stacked upwards, and then the partial stacking planning on the containers is started;
since the part of article bags on the container can extend out of the compartment, the number of article bags rows _ out = round ((W +2 extended)/(L-overlap)) which can be stacked in each row in the part of the container, the number of article bags cols _ out = round (L/(W-overlap/2)) which can be stacked in each column, the number of layers _ out = floor ((N-N _ in)/(rows _ out cols _ out)) which can be stacked in the part of the container, floor is a rounded-down function, and the remainder is represented by c. The stacking bag number of each layer on the upper part of the container is rows _ out _ cols _ out, the threshold range of the remainder c of the layers _ out is set to be [ alpha (rows _ out _ cols _ out) and beta (rows _ out _ cols _ out) ], and if the remainder c falls into the intervals [ alpha (rows _ out _ cols _ out) and beta (rows _ out _ cols _ out) ], the remainder can realize top-layer cross-stitch stacking, namely the remainder layer is the top layer; if the remainder c is smaller than alpha (rows _ out _ cols _ out) or larger than beta (rows _ out _ cols _ out), the number of the stacking bags on the uppermost layer of the goods is too large or too small, and in order to ensure the stacking stability, the values of the extension and the overlap on the upper part of the packing box are adjusted, so that the remainder value is adjusted to fall into the intervals of alpha (rows _ out _ cols _ out) and beta (rows _ out _ cols _ out);
the total number of bags stacked on the part of the container, wherein N _ out = layers _ out _ cols _ out _ rows _ out;
the remainder c is the number N _ top of stacking bags on the top layer of the container;
f. the top layer part needs to be stacked on seams for ensuring stacking stability, so that the number of bags which can be stacked in each row of the top layer part is rows _ top = round ((W + 2. extend-L)/(L-overlap)), and the number of bags which can be stacked in each column is cols _ top = round (L/(W-overlap/2));
g. and (4) knowing the stacking planning result of each layer, sequentially extracting the coordinates of each bag according to the size of the bagged articles and the coordinates of the truck, and completing the stacking planning.
In another aspect of the invention, there is also provided an open container loading three-dimensional sensing and stacking planning system, comprising;
a sensing unit: carrying a 2D laser scanning radar on a transverse moving synchronous belt to move along with the synchronous belt to serve as a sensing unit, and collecting three-dimensional laser point cloud data;
the automatic calibration unit is used for automatically calibrating to obtain a coordinate conversion relation between a sensing unit coordinate system and a loading unit coordinate system;
the background modeling and filtering unit is used for modeling the background, filtering based on a background model and filtering background points;
the data preprocessing unit is used for carrying out noise reduction processing on the point cloud data processed by the background modeling and filtering unit;
the container detection and pose estimation unit is used for carrying out container detection and pose estimation according to the point cloud data preprocessed by the data preprocessing unit;
and the stacking planning unit is used for performing stacking planning according to the size and the pose of the container obtained by the container detection and pose estimation unit and according to the size information of the known goods.
Further, the automatic calibration unit comprises:
an acquisition module: after the artificial marker is installed on the fixed object in the movement range of the loading unit, the artificial marker is used for instructing the loading unit to move to the position of the artificial marker in the loading operation space, and respectively collecting a coordinate P _ sense of the artificial marker in a sensing unit coordinate system and a coordinate P _ load in a loading unit coordinate system;
a calibration module: constructing a homogeneous coordinate matrix P _ s _ matrix of P _ sense; and (3) constructing a coordinate matrix P _ l _ matrix of the P _ load, calculating a generalized inverse matrix P _ s _ matrix _ inv of the P _ s _ matrix, calculating a transformation matrix of a sensing unit coordinate system and a loading unit coordinate system as [ R t ] = P _ l _ matrix x P _ s _ matrix _ inv, calculating a least square optimal coordinate system transformation matrix, and completing automatic calibration.
Further, the background modeling and filtering unit includes:
background offline modeling module: the system comprises a background model configuration file, a region growing algorithm, a three-dimensional laser point cloud data acquisition module, a region growing algorithm module, a background model configuration file and a data processing module, wherein the three-dimensional laser point cloud data acquisition module is used for acquiring three-dimensional laser point cloud data of a background, calculating a vertical normal vector and a curvature of each point of the point cloud, dividing the background point cloud data into a plurality of surface elements by using the region growing algorithm, calculating a central point, a normal direction and a maximum radius of each surface element as characteristics, and sequentially storing;
and the background online filtering module is used for loading a background model, calculating the distance from each point in the current point cloud to each background surface element after point cloud data containing the open container truck is obtained, so as to judge whether the point is in the size range of the background surface element, judging whether the current point belongs to the background according to a specific threshold value, filtering the point if the point belongs to the background, and circulating until all the points finish background filtering.
Further, the container detection and pose estimation unit includes:
the segmentation module is used for segmenting the point cloud of the container by using a region growing algorithm to obtain each surface element of the open container;
the size calculation module is used for calculating the geometric size information of each surface element by using a minimum bounding box algorithm;
and the pose calculation module is used for determining coordinates of each vertex of the container in the coordinate system of the sensing unit according to the size of the open container so as to calculate the pose of the open container.
Further, the palletizing planning unit comprises:
a bag number determination module: the method is used for determining the total bag number of loading according to the total weight of loading and the weight of each bag of bagged powdery articles: n = Weight/Weight, before palletizing planning, N is a known determined value;
a distance determination module: the distance overlap used for setting the article bag and the distance extend by which the side edge of the article bag can extend out of the carriage;
a layering module: the stacking plan is divided into a container inner part, a container upper part and a top layer part from bottom to top, and article bags in the container inner part cannot be stretched out, so that the extended = 0; the top layer part is used for ensuring stability and needs to be stacked on a seam;
a cargo box interior planning module: the stacking is divided into three modules: in the container, the related parameters are expressed by in; an upper part of the cargo box, and relevant parameters are expressed by out; the top layer of the container, and the related parameters are represented by top; rows represent rows, cols represent columns, and layers represent palletizing layer numbers;
n represents the total number of bags needing loading, and N _ in represents the number of bags which can be stacked in the container; the cargo box internal planning module is used for planning stacking in an integer linear way, firstly planning the internal part of the cargo box, calculating the number of goods bags which can be stacked in each row, namely cols _ in = round (L/(w-overlap/2)), L is the length of the carriage, w is the width of the goods bag, and round (x) is a rounded integer function;
calculating the number rows _ in = round (W/(l-overlap)), where each row in the container can be stacked, W is the width of the carriage, and l is the length of the article bag;
calculating the number of layers which can be stacked in the container, wherein layers _ in = round (H/H), H is the height of the compartment, and H is the height of the goods bag;
calculating the total number of bags which can be stacked in the container, wherein N _ in = layers _ in _ cols _ in _ rows _ in;
a cargo box upper portion planning module: if N _ in < N, the goods are not completely loaded, and the goods are continuously stacked upwards, and then the partial stacking planning on the containers is started;
since the part of article bags on the container can extend out of the compartment, the number of article bags rows _ out = round ((W +2 extended)/(L-overlap)) which can be stacked in each row in the part of the container, the number of article bags cols _ out = round (L/(W-overlap/2)) which can be stacked in each column, the number of layers _ out = floor ((N-N _ in)/(rows _ out cols _ out)) which can be stacked in the part of the container, floor is a rounded-down function, and the remainder is represented by c. The stacking bag number of each layer on the upper part of the container is rows _ out _ cols _ out, the threshold range of the remainder c of the layers _ out is set to be [ alpha (rows _ out _ cols _ out) and beta (rows _ out _ cols _ out) ], and if the remainder c falls into the intervals [ alpha (rows _ out _ cols _ out) and beta (rows _ out _ cols _ out) ], the remainder can realize top-layer cross-stitch stacking, namely the remainder layer is the top layer; if the remainder c is smaller than alpha (rows _ out _ cols _ out) or larger than beta (rows _ out _ cols _ out), the number of the stacking bags on the uppermost layer of the goods is too large or too small, and in order to ensure the stacking stability, the values of the extension and the overlap on the upper part of the packing box are adjusted, so that the remainder value is adjusted to fall into the intervals of alpha (rows _ out _ cols _ out) and beta (rows _ out _ cols _ out);
the total number of bags stacked on the part of the container, wherein N _ out = layers _ out _ cols _ out _ rows _ out;
the remainder c is the number N _ top of stacking bags on the top layer of the container;
the top layer planning module of the container: for ensuring stacking stability, the bag stacking on the seams is needed, so that the number of bags which can be stacked in each row of the top layer part is rows _ top = round ((W + 2. extending-L)/(L-overlap)), and the number of bags which can be stacked in each column is cols _ top = round (L/(W-overlap/2));
a coordinate extraction module: and (4) sequentially extracting the coordinates of each bag according to the size of the bagged articles and the coordinates of the truck to finish the stacking planning by using the known stacking planning result of each layer.
Compared with the prior art, the invention has the following beneficial effects:
(1) according to the invention, the three-dimensional sensing unit is constructed by using the synchronous belt motor and the 2D laser scanning radar, so that the hardware cost is saved, and the measurement precision and resolution are improved;
(2) the invention adopts the open container detection method based on background filtering, and does not need to construct a model of the open container, so that the open container detection method has good adaptability to open containers of different types and sizes;
(3) the automatic stacking planning method based on the optimization principle and the stable loading experience realizes automatic stacking planning, further improves the intellectualization and automation level of automatic loading equipment, improves the operation efficiency and quality, and saves the loading cost of manufacturers;
(4) the invention can be widely applied to the automatic loading equipment of the open type container.
Detailed Description
The three-dimensional sensing and stacking planning method provided by the invention aims at different types of open container trucks and mainly comprises the technical schemes of three-dimensional laser point cloud data acquisition, size, position and posture estimation of the open container, a stacking planning method, automatic calibration of a sensing unit and an automatic truck loading unit, three-dimensional visualization of loading and the like. The three-dimensional sensing device has the characteristics of high precision and high speed, and the stacking planning supports various stacking modes such as pattern stacking and the like, and the two modes can run on line.
In the invention, due to the comprehensive consideration of cost and sensing precision, a three-dimensional sensing scheme that a 2D laser scanning radar is carried on a transverse moving synchronous belt and moves along with the synchronous belt is adopted, and compared with a method for rotationally scanning the 2D laser scanning radar, the method has the advantages that the measurement error in the transverse moving direction is a linear error, the phenomenon that the measurement error is increased along with the increase of the observation distance does not exist, and the measurement precision can be effectively improved; compared with the method for directly measuring by using the 3D laser scanning radar, the resolution (800 lines) and the precision of the method in the transverse moving direction are far higher than those of the existing 3D laser scanning radar (16/32 lines), and the cost is much lower.
In the actual operation process, the computer sends the motion control instruction to the motor controller through the serial ports, and the motor hold-in range drives 2D laser scanning radar along the motion of hold-in range direction, and meanwhile, 2D laser scanning radar scans along the direction of perpendicular to hold-in range. And carrying out time synchronization on the synchronous belt position data and the 2D laser scanning radar scanning data, and fusing to obtain the three-dimensional point cloud.
The laser sensor adopted in the implementation is a Sick LMS11 series 2D laser scanning radar, the angular resolution of laser is set to be 0.25 degrees, the scanning range is-45-225 degrees, the laser scanning frequency is 25HZ, the length of a transverse guide rail is 3400mm, the transverse moving speed is 80.0mm/s, the single scanning time is 42.229s, three-dimensional point cloud data including a truck and a background can be obtained, and 1099377 point cloud data are counted;
in the invention, the three-dimensional perception of the pose and the size of the open container and the stacking planning comprise the following steps:
1. and (4) automatic calibration. In order to complete accurate visual servo motion control, the invention provides an automatic calibration method for the accurate conversion relation between a sensing unit coordinate system and a loading unit coordinate system. Firstly, an artificial marker is installed on a fixed object in the movement range of a loading unit, then the loading unit is controlled to move to the artificial marker in a loading operation space through an instruction, the coordinates of the artificial marker in a sensing unit coordinate system and the coordinates of the artificial marker in a loading unit coordinate system, P _ sense and P _ load are respectively collected, and M groups of data (M > 3) are collected. And the coordinate conversion relation between the sensing unit coordinate system and the loading unit coordinate system is P _ load = [ R t ] P _ sense. Constructing a homogeneous coordinate matrix P _ s _ matrix of P _ sense; and (3) constructing a coordinate matrix P _ l _ matrix of the P _ load, calculating a generalized inverse matrix P _ s _ matrix _ inv of the P _ s _ matrix, wherein a conversion matrix of a sensing unit coordinate system and a loading unit coordinate system is [ R t ] = P _ l _ matrix x P _ s _ matrix _ inv, calculating an optimal coordinate system conversion matrix by using a least square method, and completing automatic calibration.
The method specifically comprises the following steps: acquiring point cloud data of 4 groups of markers by using laser, and obtaining a homogeneous coordinate matrix P _ s _ matrix of the markers in a sensing unit coordinate system as follows:
the instruction is utilized to control the loading unit to collect 4 groups of data of the marker at different positions, and a coordinate matrix P _ l _ matrix of the marker in a coordinate system of the loading unit is obtained as follows:
the generalized inverse matrix P _ s _ matrix _ inv of P _ s _ matrix is obtained as:
according to the formula [ R t ] = P _ l _ matrix × P _ s _ matrix _ inv, a conversion matrix [ R t ] between the sensing unit coordinate system and the loading unit coordinate system is obtained as follows:
the automatic calibration is completed.
2. Background modeling and filtering. There are many object detection, identification and segmentation algorithms based on three-dimensional point clouds. The direct detection method needs to realize modeling of the detected object and requires the detected object to have obvious characteristics. In the application scene, all models of the open container trucks cannot be obtained, and the open container trucks are generally similar and have no obvious distinguishable characteristics, so that the method is not suitable for direct detection. The invention provides an open container detection method based on background model filtering aiming at a specific application scene, which comprises the following specific steps:
2.1 background offline modeling
Firstly, three-dimensional laser point cloud data of a background is collected, and a plurality of pieces of dense laser point cloud background data can be collected and fused into one piece. And then calculating the vertical normal vector and curvature of each point of the point cloud, and segmenting the background point cloud data into a plurality of surface elements by using a region growing algorithm, wherein each surface element has similar vertical finding and curvature. Finally, calculating the central point, the normal direction and the maximum radius of each surface element as characteristics, and sequentially storing the characteristics into a background model configuration file to complete background modeling;
2.2 background on-line Filtering
Firstly loading a background model, after point cloud data containing an open container truck is obtained, calculating the distance from each point in the current point cloud to each background surface element and judging whether the change point is in the size range of the background surface element, then judging whether the current point belongs to the background according to a specific threshold value, filtering the point if the change point belongs to the background, and circulating until all the points finish background filtering.
3. And (4) preprocessing data. The point cloud after background filtering comprises the point cloud of the open container and other noise point clouds, and noise reduction processing is performed in order to improve the efficiency and the precision of a subsequent perception algorithm. Firstly, describing the spatial distribution condition of point cloud by using a Gaussian mixture model, and calculating the variance Sigma of the sparsity degree of the reaction point cloud; then setting a threshold value, and removing points larger than the threshold value; and finally, carrying out down-sampling on the new point cloud to finish data preprocessing.
4. Container detection and pose estimation. Similar to background offline modeling, the method firstly uses a region growing algorithm to segment the point cloud of the container to obtain each surface element of the open container; then, calculating the geometric dimension information of each surface element by using a minimum bounding box algorithm, wherein the main idea of the minimum bounding box algorithm is to perform Principal Component Analysis (PCA) on a covariance matrix of the point cloud to determine principal components of the point cloud in the three-dimensional direction, so that a minimum bounding box is generated; finally, calculating the mutual position relation of the set surface elements according to the geometrical relation of the open type container surface elements, such as the constraint that the area of the bottom surface is maximum, the long edge of the bottom surface is collinear with the left side surface and the right side surface, the short edge of the bottom surface is collinear with the front side surface and the rear side surface, and the like; and then determining the weight of each surface element according to the point cloud sparsity and the observation state of each surface element, and deducing the length, width, height and other dimensions of the open container. And finally, determining coordinates of each vertex of the container in a coordinate system of the sensing unit according to the size of the open container, so as to calculate the pose of the container. Substituting the filtered point cloud data into a three-dimensional perception algorithm, and estimating the size of the boxcar [4202,2365 and 429], wherein the real size of the boxcar [4200,2350,410] mm; the errors of the three dimensions are all less than 20 mm; the position of the truck vehicle is [ -527,363, -3109] mm, and the three positions and postures are [0.9999, -0.0044,0.0013], [0.0043,0.9999, -0.0009], [ -0.0014,0.0008 and 0.9999 ].
5. And (6) palletizing planning. Generally, goods to be packaged are standard sizes meeting industry standards, such as cement, white granulated sugar, flour and the like, so that the size information of the goods is known information. Dividing the stacking plan into a container inner part, a container upper part and a top layer part from bottom to top, wherein the container inner part and the container upper part adopt integer linear plan, and the top layer part is required to be subjected to seam riding stacking for ensuring stability;
the specific implementation mode is as follows: assuming that 10 tons of articles are loaded, the weight of each bag of articles is 50kg, the size of each article bag is [600,400,200] mm, the overlapping distance of the article bags during stacking is set to be 100mm, the distance of each article bag side edge capable of extending out of a carriage is set to be 100mm, the size of a vehicle body is sensed to be [4202,2365, 429] mm, the total number N of bags for loading is determined to be 10000/50=200 bags according to a formula, then stacking planning is carried out, the inner part of a container is planned, the number of article bags which can be stacked in each row is calculated to be cols _ in = round (L/(w-overlap/2)), L is the length of the carriage, w is the width of the article bag, round (×) is a rounding function, and the value is substituted to be cols _ in = (round (4202/(400-50)) = 12); calculating the number of article bags capable of being stacked in each row, namely rows _ in = round (W/(l-overlap)), W is the width of the carriage, l is the length of the article bag, and substituting the value into rows _ in = (round (2365/(600-; calculating the number of layers which can be stacked in the container, wherein layer _ in = round (H/H), H is the height of the compartment, H is the height of the goods bag, and the value is substituted into layer _ in = round (429/200) = 2; calculating the total number of bags which can be stacked in the container, wherein N _ in = layers _ in _ cols _ in _ rows _ in =2 12 = 5= 120; the upper part of the compartment is planned continuously due to N _ in < N, and the number of bags capable of being stacked in each row, namely, rows _ out = round ((W +2 + extend)/(L-overlap)) = round ((2365+2 + 100)/(600 + 100)) =5; cols _ out = round (L/(W-overlap/2)) = (round (4202/(400-50))) =12; layers _ out = floor ((N-N _ in)/(rows _ cols _ out)) = (200 + 120)/(12 + 5) =1 … 20) as the bags in the compartment can extend out of the compartment; calculating the total number of bags which can be stacked on the container, wherein N _ out = layers _ out _ colors _ out _ rows _ out =1 _ 12 _ 5=60, the remainder =20, N _ top =20, the top layer part needs to be subjected to seam-crossing stacking for ensuring stacking stability, so that the number of bags which can be stacked on each row is colors _ top = (W +2 _ extended-l)/(l-overlap) = (2365+ 2-; the palletizing planning result is:
a. 120 bags are loaded in the carriage, 2 layers and 12 rows and 5 rows
b. 60 bags are loaded outside the carriage, and 1 layer, 12 rows and 5 rows
c. The top layer is loaded with 20 bags, 1 layer, 4 columns and 5 rows
6. And (4) three-dimensional visualization. After the size, the pose information and the stacking planning structure of the open type container are obtained, the open type container can be displayed in a three-dimensional visual interface according to the real size, and the loading result of the goods is sequentially displayed according to the stacking planning result.
7. And (5) visual servo loading operation. The motion control of the article loading unit is controlled by a PLC, the results of three-dimensional sensing and stacking planning are converted to the position below a loading unit coordinate system and are sent to a PLC controller through a communication protocol specific to the PLC, the PLC controller controls the loading mechanism to start stacking after receiving the results, and the state of the PLC controller can be inquired in real time in the process.
The invention provides a three-dimensional sensing method for the position, posture and size of an open container and a bagged object stacking planning method, which can enable automatic truck loading equipment to automatically sense different poses and different truck types when trucks are parked, automatically plan a stacking and loading method according to the loading amount, and visually servo-control the truck loading equipment to finish automatic truck loading, so that the whole truck loading process does not need manual intervention completely, the automation level and efficiency of open container loading are greatly improved, and the cost is saved for manufacturers.
The above description is only a preferred embodiment of the present invention, and it should be noted that, for those skilled in the art, several modifications and decorations can be made without departing from the concept of the present invention, and these modifications and decorations should also be regarded as being within the protection scope of the present invention.
Claims (8)
1. The utility model provides an open packing box loading three-dimensional perception and pile up neatly planning method which characterized in that carries on 2D laser scanning radar on the sideslip hold-in range and removes as the perception unit collection three-dimensional laser point cloud data along with the hold-in range, still includes:
s1, automatically calibrating to obtain a coordinate conversion relation between a sensing unit coordinate system and a loading unit coordinate system;
s2, background modeling, filtering based on the background model, and filtering background points;
s3, preprocessing data, and performing noise reduction processing on the point cloud data obtained in the step S2;
s4, carrying out container size detection and pose estimation according to the preprocessed point cloud data;
s5, carrying out stacking planning according to the size and the pose of the cargo box obtained in the S4 and the size information of the known cargo;
in step S2, the method specifically includes:
s21, background offline modeling: acquiring three-dimensional laser point cloud data of a background, calculating a vertical normal vector and curvature of each point of the point cloud, dividing the background point cloud data into a plurality of surface elements by using a region growing algorithm, wherein each surface element has similar vertical normal and curvature, calculating a central point, a normal direction and a maximum radius of each surface element as characteristics, and sequentially storing the characteristics in a background model configuration file to complete background modeling;
s22, background online filtering: loading a background model; after point cloud data containing the open container truck is obtained, the distance from each point in the current point cloud to each background surface element is calculated to judge whether the point is in the size range of the background surface element, then whether the current point belongs to the background is judged according to a specific threshold value, if the current point belongs to the background, the point is filtered, and the process is repeated until all the points complete background filtering.
2. The method according to claim 1, wherein the automatic calibration method of step S1 includes:
s11, mounting an artificial marker on a fixed object in the motion range of the loading unit;
s12, controlling the loading unit to move to the position of the artificial marker in the loading operation space through an instruction, and respectively acquiring a coordinate P _ sense of the artificial marker in a sensing unit coordinate system and a coordinate P _ load in a loading unit coordinate system;
s13, constructing a homogeneous coordinate matrix P _ S _ matrix of P _ sense; and (3) constructing a coordinate matrix P _ l _ matrix of the P _ load, calculating a generalized inverse matrix P _ s _ matrix _ inv of the P _ s _ matrix, wherein a conversion matrix of a sensing unit coordinate system and a loading unit coordinate system is (R t) ═ P _ l _ matrix × P _ s _ matrix _ inv, calculating an optimal coordinate system conversion matrix by using a least square method, and completing automatic calibration.
3. The method of claim 1, wherein the specific method of container detection and pose estimation in step S4 comprises:
s41, segmenting the point cloud of the container by using a region growing algorithm to obtain each surface element of the open container;
s42, calculating the geometric dimension information of each surface element by using a minimum bounding box algorithm;
and S43, determining coordinates of each vertex of the container in the coordinate system of the sensing unit according to the size of the open container, and calculating the pose of the open container.
4. The method of claim 1, wherein the specific method of palletizing planning in step S5 comprises:
a. calculating the total bag number of the loading according to the total weight of the loading and the weight of each bag of the bagged powdery materials: n is Weight/Weight;
b. determining the overlapping distance of the article bags and the distance extended by which the side edge of the article bag can extend out of the carriage;
c. the stacking plan is divided into a container inner part, a container upper part and a top layer part from bottom to top, and article bags in the container inner part cannot be stretched out, so that the extension is 0; the top layer part is used for ensuring stability and needs to be stacked on a seam;
d. the stacking is divided into three modules: in the container, the related parameters are expressed by in; an upper part of the cargo box, and relevant parameters are expressed by out; the top layer of the container, and the related parameters are represented by top; rows represent rows, cols represent columns, and layers represent palletizing layer numbers;
n represents the total number of bags needing loading, and N _ in represents the number of bags which can be stacked in the container;
the stacking planning adopts integer linear planning, firstly, part of a container is planned, the number of goods bags which can be stacked in each row in the container is calculated, namely cols _ in is round (L/(w-overlap/2)), L is the length of a carriage, w is the width of the goods bag, and round (x) is a rounding function;
calculating the number of the goods bags capable of being stacked in each row in the container, namely row _ in, wherein W is the width of the carriage, and l is the length of the goods bag;
calculating the number of layers which can be stacked in the container, wherein layers _ in is round (H/H), H is the height of the carriage, and H is the height of the goods bag;
calculating the total number of bags which can be stacked in the container, wherein N _ in is layers _ in _ cols _ in rows _ in;
e. if N _ in < N, the goods are not completely loaded, and the goods are continuously stacked upwards, and then the partial stacking planning on the containers is started;
since the partial article bags on the container can extend out of the compartment, the number of article bags rows _ out which can be stacked in each row on the container is round ((W +2 extended)/(L-overlap)), the number of article bags cols _ out which can be stacked in each column is round (L/(W-overlap/2)), the number of layers _ out which can be stacked on the container is floor ((N-N _ in)/(rows _ out) cols _ out)), floor is a rounding-down function, and the remainder is represented by c; the stacking bag number of each layer on the upper part of the container is rows _ out _ cols _ out, the threshold range of the remainder c of the layers _ out is set to be [ alpha (rows _ out _ cols _ out) and beta (rows _ out _ cols _ out) ], and if the remainder c falls into the intervals [ alpha (rows _ out _ cols _ out) and beta (rows _ out _ cols _ out) ], the remainder can realize top-layer cross-stitch stacking, namely the remainder layer is the top layer; if the remainder c is smaller than alpha (rows _ out _ cols _ out) or larger than beta (rows _ out _ cols _ out), the number of the stacking bags on the uppermost layer of the goods is too large or too small, and in order to ensure the stacking stability, the values of the extension and the overlap on the upper part of the packing box are adjusted, so that the remainder value is adjusted to fall into the intervals of alpha (rows _ out _ cols _ out) and beta (rows _ out _ cols _ out);
the total bag number of the part of stacking on the container, N _ out is layers _ out _ cols _ out _ rows _ out;
the remainder c is the number N _ top of stacking bags on the top layer of the container;
f. the top layer part needs to be stacked on seams for ensuring stacking stability, so that the number of bags which can be stacked in each row of the top layer part is rows _ top ═ round ((W +2 × extended-L)/(L-overlap)), and the number of bags which can be stacked in each column is colors _ top ═ round (L/(W-overlap/2));
g. and (4) knowing the stacking planning result of each layer, sequentially extracting the coordinates of each bag according to the size of the bagged articles and the coordinates of the truck, and completing the stacking planning.
5. The utility model provides an open packing box loading three-dimensional perception and pile up neatly planning system which characterized in that includes:
a sensing unit: carrying a 2D laser scanning radar on a transverse moving synchronous belt to move along with the synchronous belt to serve as a sensing unit, and collecting three-dimensional laser point cloud data;
the automatic calibration unit is used for automatically calibrating to obtain a coordinate conversion relation between a sensing unit coordinate system and a loading unit coordinate system;
the background modeling and filtering unit is used for modeling the background, filtering based on a background model and filtering background points;
the data preprocessing unit is used for carrying out noise reduction processing on the point cloud data processed by the background modeling and filtering unit;
the container detection and pose estimation unit is used for carrying out container detection and pose estimation according to the point cloud data preprocessed by the data preprocessing unit;
the stacking planning unit is used for performing stacking planning according to the size and the pose of the container obtained by the container detection and pose estimation unit and according to the size information of the known goods;
the background modeling and filtering unit includes:
background offline modeling module: the system comprises a background model configuration file, a region growing algorithm, a three-dimensional laser point cloud data acquisition module, a region growing algorithm module, a background model configuration file and a data processing module, wherein the three-dimensional laser point cloud data acquisition module is used for acquiring three-dimensional laser point cloud data of a background, calculating a vertical normal vector and a curvature of each point of the point cloud, dividing the background point cloud data into a plurality of surface elements by using the region growing algorithm, calculating a central point, a normal direction and a maximum radius of each surface element as characteristics, and sequentially storing;
and the background online filtering module is used for loading a background model, calculating the distance from each point in the current point cloud to each background surface element after point cloud data containing the open container truck is obtained, so as to judge whether the point is in the size range of the background surface element, judging whether the current point belongs to the background according to a specific threshold value, filtering the point if the point belongs to the background, and circulating until all the points finish background filtering.
6. The system of claim 5, wherein the automated calibration unit comprises:
an acquisition module: after the artificial marker is installed on the fixed object in the movement range of the loading unit, the artificial marker is used for instructing the loading unit to move to the position of the artificial marker in the loading operation space, and respectively collecting a coordinate P _ sense of the artificial marker in a sensing unit coordinate system and a coordinate P _ load in a loading unit coordinate system;
a calibration module: constructing a homogeneous coordinate matrix P _ s _ matrix of P _ sense; and (3) constructing a coordinate matrix P _ l _ matrix of the P _ load, calculating a generalized inverse matrix P _ s _ matrix _ inv of the P _ s _ matrix, calculating a transformation matrix of a sensing unit coordinate system and a loading unit coordinate system as [ R t ] ═ P _ l _ matrix x P _ s _ matrix _ inv, calculating a least square optimal coordinate system transformation matrix, and finishing automatic calibration.
7. The system of claim 5, wherein the container detection and pose estimation unit comprises:
the segmentation module is used for segmenting the point cloud of the container by using a region growing algorithm to obtain each surface element of the open container;
the size calculation module is used for calculating the geometric size information of each surface element by using a minimum bounding box algorithm;
and the pose calculation module is used for determining coordinates of each vertex of the container in the coordinate system of the sensing unit according to the size of the open container so as to calculate the pose of the open container.
8. The system of claim 5, wherein the palletization planning unit comprises:
a bag number determination module: the method is used for determining the total bag number of loading according to the total weight of loading and the weight of each bag of bagged powdery articles: n is known determined value before palletizing planning;
a distance determination module: the distance overlap used for setting the article bag and the distance extend by which the side edge of the article bag can extend out of the carriage;
a layering module: the stacking plan is divided into a container inner part, a container upper part and a top layer part from bottom to top, and article bags in the container inner part cannot be extended out, so that the extension is 0; the top layer part is used for ensuring stability and needs to be stacked on a seam;
a cargo box interior planning module: the stacking is divided into three modules: in the container, the related parameters are expressed by in; an upper part of the cargo box, and relevant parameters are expressed by out; the top layer of the container, and the related parameters are represented by top; rows represent rows, cols represent columns, and layers represent palletizing layer numbers;
n represents the total number of bags needing loading, and N _ in represents the number of bags which can be stacked in the container; the cargo box internal planning module is used for planning stacking planning by adopting integer linear planning, firstly planning the internal part of a cargo box, and calculating the number cols _ in of each row of stackable object bags as round (L/(w-overlap/2)), wherein L is the length of the carriage, w is the width of the object bag, and round (x) is a rounding function;
calculating the number of the goods bags capable of being stacked in each row in the container, namely row _ in, wherein W is the width of the carriage, and l is the length of the goods bag;
calculating the number of layers which can be stacked in the container, wherein layers _ in is round (H/H), H is the height of the carriage, and H is the height of the goods bag;
calculating the total number of bags which can be stacked in the container, wherein N _ in is layers _ in _ cols _ in rows _ in;
a cargo box upper portion planning module: if N _ in < N, the goods are not completely loaded, and the goods are continuously stacked upwards, and then the partial stacking planning on the containers is started;
since the partial article bags on the container can project out of the vehicle, the number of article bags capable of being stacked in each row on the container, rows _ out, rows (L/(W-overlap/2)), the number of article bags capable of being stacked in each column, stacks _ out, floor (N-in)/(rows _ out), floor is a downward integer function, the remainder is represented by c, the number of stacked bags in each layer on the container is rows _ out, the threshold range of the remainder of layers _ out is set to [ α (rows _ out _ rows _ out), β (rows _ out _ rows _ out) ], and the remainder is realized if the c falls in the interval [ α (rows _ out) and β (rows _ out) ], namely the remainder layer is the top layer; if the remainder c is smaller than alpha (rows _ out _ cols _ out) or larger than beta (rows _ out _ cols _ out), the number of the stacking bags on the uppermost layer of the goods is too large or too small, and in order to ensure the stacking stability, the values of the extension and the overlap on the upper part of the packing box are adjusted, so that the remainder value is adjusted to fall into the intervals of alpha (rows _ out _ cols _ out) and beta (rows _ out _ cols _ out);
the total bag number of the part of stacking on the container, N _ out is layers _ out _ cols _ out _ rows _ out;
the remainder c is the number N _ top of stacking bags on the top layer of the container;
the top layer planning module of the container: for ensuring stacking stability, the bags need to be stacked on seams, so that the number of the bags which can be stacked in each row of the top layer part is rows _ top ═ round ((W +2 x extended-L)/(L-overlap)), and the number of the bags which can be stacked in each column is cols _ top ═ round (L/(W-overlap/2));
a coordinate extraction module: and (4) sequentially extracting the coordinates of each bag according to the size of the bagged articles and the coordinates of the truck to finish the stacking planning by using the known stacking planning result of each layer.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010780533.1A CN111652936B (en) | 2020-08-06 | 2020-08-06 | Three-dimensional sensing and stacking planning method and system for open container loading |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010780533.1A CN111652936B (en) | 2020-08-06 | 2020-08-06 | Three-dimensional sensing and stacking planning method and system for open container loading |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111652936A CN111652936A (en) | 2020-09-11 |
| CN111652936B true CN111652936B (en) | 2020-11-06 |
Family
ID=72351808
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010780533.1A Active CN111652936B (en) | 2020-08-06 | 2020-08-06 | Three-dimensional sensing and stacking planning method and system for open container loading |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111652936B (en) |
Families Citing this family (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112767540A (en) * | 2021-01-15 | 2021-05-07 | 苏州艾吉威机器人有限公司 | Automatic loading method for AGV |
| CN114275222A (en) * | 2021-01-19 | 2022-04-05 | 湖北省烟草公司宜昌市公司 | Automatic stacking device and method for order cigarette carton cams |
| CN115598614A (en) * | 2022-11-28 | 2023-01-13 | 南京隼眼电子科技有限公司（Cn） | Three-dimensional point cloud target detection method and device and storage medium |
| CN115984278B (en) * | 2023-03-20 | 2023-07-07 | 四川吉埃智能科技有限公司 | Loading and stacking planning method and system based on full-feature information of vehicle to be loaded |
| CN116630429B (en) * | 2023-07-26 | 2023-10-31 | 成都物天物联网科技有限责任公司 | Visual guiding and positioning method and device for docking of vehicle and box and electronic equipment |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104598675A (en) * | 2015-01-07 | 2015-05-06 | 北京卫星环境工程研究所 | Actually measured data-based spacecraft component assembly simulation method |
| US10133400B2 (en) * | 2015-07-01 | 2018-11-20 | Tactual Labs Co. | Pressure informed decimation strategies for input event processing |
Family Cites Families (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106429483A (en) * | 2016-12-09 | 2017-02-22 | 芜湖哈特机器人产业技术研究院有限公司 | Automatic stacking and truck-loading system and truck-loading method thereof |
| CN107618894B (en) * | 2017-09-27 | 2023-04-07 | 四川福德机器人股份有限公司 | Flexible automatic loading system and application thereof |
| CN107963472A (en) * | 2017-12-22 | 2018-04-27 | 长沙长泰机器人有限公司 | Robot automatic loading system and method based on laser vision |
| CN109095059B (en) * | 2018-07-06 | 2021-06-22 | 青岛智能产业技术研究院 | Unmanned loading system and loading method thereof |
| CN109399250A (en) * | 2018-11-28 | 2019-03-01 | 青岛万龙智控科技有限公司 | Lift truck calibration system and method |
| CN109720891A (en) * | 2018-12-29 | 2019-05-07 | 浙江明度智控科技有限公司 | A kind of cargo loads planning system and method automatically |
| CN110340891B (en) * | 2019-07-11 | 2022-05-24 | 河海大学常州校区 | Mechanical arm positioning and grabbing system and method based on point cloud template matching technology |
| CN111121628A (en) * | 2019-12-31 | 2020-05-08 | 芜湖哈特机器人产业技术研究院有限公司 | Calibration method of three-dimensional scanning system of carriage container based on two-dimensional laser radar |
-
2020
- 2020-08-06 CN CN202010780533.1A patent/CN111652936B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104598675A (en) * | 2015-01-07 | 2015-05-06 | 北京卫星环境工程研究所 | Actually measured data-based spacecraft component assembly simulation method |
| US10133400B2 (en) * | 2015-07-01 | 2018-11-20 | Tactual Labs Co. | Pressure informed decimation strategies for input event processing |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111652936A (en) | 2020-09-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111652936B (en) | Three-dimensional sensing and stacking planning method and system for open container loading | |
| US9630320B1 (en) | Detection and reconstruction of an environment to facilitate robotic interaction with the environment | |
| EP3169489B1 (en) | Real-time determination of object metrics for trajectory planning | |
| JP6850867B2 (en) | Automatic storage and acquisition system with detectors to detect items that extend beyond the dimensional threshold | |
| CN104085794B (en) | Intelligent detection system and method for distribution profile and position of container | |
| EP4223461A2 (en) | Robotic carton unloader with integral extraction mechanism | |
| WO2016010968A1 (en) | Multiple suction cup control | |
| CN108007451A (en) | Detection method, device, computer equipment and the storage medium of cargo carrying device pose | |
| CN112010177B (en) | Automatic container landing method for ground containers in storage yard | |
| CN112850186B (en) | Mixed pile-dismantling method based on 3D vision | |
| CN111498525A (en) | Automatic loading positioning system and method | |
| CN111192328A (en) | Two-dimensional laser radar-based point cloud processing method for three-dimensional scanning system of compartment container | |
| Bellandi et al. | Roboscan: a combined 2D and 3D vision system for improved speed and flexibility in pick-and-place operation | |
| CN106379685A (en) | Automatic goods taking and unloading equipment based on unmanned carrier | |
| CN115328175A (en) | Logistics robot system | |
| CN113050636A (en) | Control method, system and device for autonomous tray picking of forklift | |
| CN112278891B (en) | Carriage internal attitude detection method | |
| CN113819844A (en) | Method, device and equipment for identifying traveling materials and readable storage medium | |
| CN111275753B (en) | Method for measuring volume of sand and stone in carriage | |
| CN115351814A (en) | Remove three-dimensional storage device that collects | |
| CN113463719B (en) | Autonomous operation control system and method for loader | |
| CN115468561A (en) | Navigation method of combined continuous material taking carrier for bulk material storage yard | |
| CN116342695B (en) | Unmanned forklift truck goods placing detection method and device, unmanned forklift truck and storage medium | |
| CN113740098B (en) | Bulk coal car sampling method | |
| CN115546029A (en) | Automatic loading system based on visual detection and control method thereof |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |