3D vision positioning error correction and alarm method for multi-layer shelves

Abstract

The 3D visual positioning error correction and alarm method for multi-layer shelving, described in this invention, employs a dual verification mechanism for positioning error correction. This mechanism, based on correcting the taught position, adds error correction to the visual positioning results themselves, in addition to sequentially updating and storing the previous layer's position information. This method aims to improve the accuracy and stability of 3D visual positioning, enabling timely detection and correction of positioning errors, preventing system failures caused by cumulative errors, and ensuring safety. The method includes the following implementation steps: Step 1) determining the taught position; Step 2) initial positioning; and Step 3) non-initial positioning.

Description

Technical Field

[0001] The invention relates to a 3D visual positioning error correction and alarm method for multi-layer shelves, and belongs to the field of automation control and logistics warehousing. Background Art

[0002] With the rapid development of data and electrical technology, information technology is increasingly being applied to automated equipment in industrial production and logistics warehousing. For example, currently automated warehousing systems typically utilize robotic arms, automated guided vehicles (AGVs), and rail-guided vehicles (RGVs) to store and move goods within a three-dimensional warehouse. Due to the uncertainty of shelf position movement, deformation, and environmental factors within a warehousing system, ensuring the precise positioning of robotic arms within multi-layered racks is crucial for the system's proper operation.

[0003] In existing technologies, 3D visual positioning technology has been widely used for object recognition and positioning. It uses cameras to obtain 3D information about shelves and combines it with image processing algorithms to achieve shelf positioning. However, a single 3D visual positioning method may be affected by factors such as light and occlusion in complex environments, resulting in reduced positioning accuracy and, in turn, failure to place goods. To improve the accuracy and reliability of 3D visual positioning, existing research results have proposed multi-sensor fusion solutions, such as combining multiple sensing technologies such as lidar and ultrasonic sensors to make up for the shortcomings of a single visual sensor. However, multi-sensor systems are complex and costly, and in practical applications, the calibration and data fusion algorithms between sensors still have obvious flaws.

[0004] In this regard, the prior art also discloses a verification mechanism based on the teaching position, which compares the 3D visual positioning result with the pre-set teaching position during the first positioning, in order to ensure the accuracy of the positioning. For example, in the prior patent application CN202311588458.9, the name is automatic error correction and fault-tolerant alarm method, system and device for stacker warehouse location data. This application uses visual sensors and shelf positioning holes to calibrate the reference cargo location points, thereby realizing automatic error correction and fault-tolerant alarm of warehouse location data. However, it only proposes the function of correcting the teaching position, and lacks error correction of the visual positioning results. After long-term use of facilities such as shelves and long-term operation of equipment for handling goods, it is inevitable that environmental factors will change and shelves will be deformed, and the effectiveness of the above-mentioned physical teaching position will gradually decrease. Moreover, the single verification mechanism adopted in this solution still has potential cumulative error problems for multi-layer shelf structures, resulting in a decrease in verification accuracy.

[0005] In view of this, this patent application is hereby filed. Summary of the Invention

[0006] The present application proposes a 3D visual positioning error correction and alarm method for multi-layer shelves, which aims to solve the problems existing in the above-mentioned existing technologies and adopts a double verification mechanism for positioning error correction, that is, taking the error correction of the teaching position as a prerequisite, on the basis of sequentially updating and storing the position information of the previous layer, adding error correction detection of the visual positioning result itself, in order to improve the accuracy and stability of 3D visual positioning, and achieve the purpose of timely discovering and correcting positioning errors, and preventing system failures and safety due to cumulative errors.

[0007] To achieve the above design objectives, the 3D visual positioning error correction and alarm method for multi-layer shelves includes the following implementation steps:

[0008] Step 1) Determine the teaching position;

[0009] Before the goods are put in or taken out of the warehouse, the teaching position of the robot arm for each cargo location is marked;

[0010] Step 2), first positioning;

[0011] Before using the robotic arm to perform a loading and unloading operation at a certain cargo location, it takes a photo with a 3D camera to obtain the 3D positioning result coordinates of the current cargo location. The 3D positioning result coordinates are compared with the taught position of the current cargo location. If the comparison difference does not exceed the preset threshold, the robotic arm performs the loading and unloading operation and uses the 3D positioning result coordinates of the current cargo location as the new taught position information and updates it to the system database. If the comparison difference exceeds the preset threshold, an alarm signal is triggered and output to prompt the operator to manually intervene and calibrate.

[0012] Step 3) Non-first positioning;

[0013] First, determine whether the current shelf is on the bottom layer. If so, perform a difference calculation between the 3D positioning result coordinates of the current shelf and the teaching position of the shelf on its left bottom layer stored in the system database. If not, perform a difference calculation between the 3D positioning result coordinates of the current shelf and the teaching position of the shelf on the next layer stored in the system database.

[0014] Secondly, it is determined whether the difference calculation results in any of the above two situations exceed the preset threshold. If not, the robot arm performs the goods in and out of the warehouse and updates the 3D positioning results of the current cargo location as the new teaching position information to the system database. If so, the controller performs a difference calculation between the 3D positioning result coordinates of the current cargo location and the teaching position of the current cargo location stored in the system database.

[0015] Finally, it is determined whether the difference between the 3D positioning result coordinates of the current cargo location and its taught position exceeds the preset threshold; if not, the robotic arm performs the goods in and out of the warehouse, and updates the 3D positioning result of the current cargo location as the new taught position information to the system database; if so, an alarm signal is triggered and output to prompt the operator to perform manual intervention and calibration.

[0016] Furthermore, the step 1) includes the following steps:

[0017] Step 1.1, mark the teaching position of each cargo location;

[0018] For the bottom shelf of each row of shelves, the teaching position is manually measured and marked; for the other shelves in the same row, that is, the shelf height data of the other shelves other than the bottom shelf is accumulated layer by layer from the teaching position of the bottom shelf.

[0019] Step 1.2, establish the three-dimensional coordinate system of the shelf;

[0020] Use a 3D camera to take pictures to obtain and establish the three-dimensional coordinate system where the multi-layer shelves are located.

[0021] Furthermore, the current 3D positioning result coordinates (X, Y, Z, W, P, R) of any cargo location are identified and calculated, where X, Y, and Z represent the X-axis coordinate, Y-axis coordinate, and Z-axis coordinate in the three-dimensional coordinate system, respectively; W is the angle of rotation around the X-axis, P is the angle of rotation around the Y-axis, and R is the angle of rotation around the Z-axis.

[0022] Furthermore, the step 2) includes the following steps:

[0023] Step 2.1, determine whether it is the first positioning;

[0024] If the current cargo location is the first time to carry out the storage and outgoing operation, it is the first positioning;

[0025] If the current location is the bottom shelf on the leftmost side of the entire shelf, it is the first positioning; otherwise, it is not the first positioning and jumps directly to step 3);

[0026] Step 2.2, perform positioning information comparison;

[0027] The 3D camera takes a picture to obtain the point cloud data of the current cargo location, and the controller calculates the 3D positioning result coordinates (X, Y, Z, W, P, R) of the current cargo location and the teaching position information (X n ,Y n ,Z n ,W n ,P n ,R n ) performs a difference operation;

[0028] Difference 1 = |XX n |, difference 2 = |YY n |, difference 3 = |ZZ n |;

[0029] Difference 4 = |WW n |, difference 5 = |PP n |, difference 6 = |RR n |;

[0030] Determine whether any of the above difference values ​​1 to 6 exceeds the preset threshold; if not, the robotic arm performs the goods in and out of the warehouse, and updates the 3D positioning result of the current cargo location as the new teaching position information to the system database; if so, trigger and output an alarm signal to prompt the operator to manually intervene and calibrate.

[0031] Furthermore, the step 3) includes the following steps:

[0032] Step 3.1, positioning of the current cargo location;

[0033] The controller determines whether the current cargo location is at the bottom layer based on the received current cargo location number;

[0034] If so, perform difference calculation on the 3D positioning result coordinates (X, Y, Z, W, P, R) of the current cargo location and the teaching position (X1, Y1, Z1, W1, P1, R1) of the bottom cargo location on its left stored in the system database;

[0035] Difference 1 = |X - X1|, Difference 2 = |Y - Y1|, Difference 3 = |Z - Z1|;

[0036] Difference 4 = |W-W1|, Difference 5 = |P-P1|, Difference 6 = |R-R1|;

[0037] If not, the 3D positioning result coordinates (X, Y, Z, W, P, R) of the current cargo location are combined with the teaching position (X, Y, Z, W, P, R) of the next cargo location stored in the system database. n-1 ,Y n-1 ,Z n-1 ,W n-1 ,P n-1 ,R n-1 ) performs a difference operation;

[0038] Difference 1 = |XX n-1 |, difference 2 = |YY n-1 |, difference 3 = |ZZ n-1 |;

[0039] Difference 4 = |WW n-1 |, difference 5 = |PP n-1|, difference 6 = |RR n-1 |;

[0040] Step 3.2: Determine whether the difference calculation result in any of the above two situations exceeds a preset threshold;

[0041] If any of the difference values ​​1 to 6 in the above two situations do not exceed the preset threshold, the robot arm will perform the goods in and out of the warehouse and update the 3D positioning result (X, Y, Z, W, P, R) of the current cargo location as the new teaching position information to the system database;

[0042] If any of the difference values ​​1 to 6 in the above two situations exceeds the preset threshold, the process proceeds to step 3.3;

[0043] Step 3.3: Calculate the difference between the teaching position information and make a judgment;

[0044] Determine the 3D positioning result coordinates (X, Y, Z, W, P, R) of the current cargo location and the teaching position information (X n ,Y n ,Z n ,W n ,P n ,R n ) performs a difference operation;

[0045] Difference 1 = |XX n |, difference 2 = |YY n |, difference 3 = |ZZ n |;

[0046] Difference 4 = |WW n |, difference 5 = |PP n |, difference 6 = |RR n |;

[0047] Determine whether any difference calculation result among the above result differences 1 to 6 exceeds a preset threshold;

[0048] If not, the robot arm performs the goods in and out of the warehouse, and updates the 3D positioning result (X, Y, Z, W, P, R) of the current cargo location as the new teaching position information to the system database;

[0049] If so, an alarm signal is triggered and output to prompt the operator to perform manual intervention and calibration.

[0050] In summary, the 3D vision positioning error correction and alarm method for multi-layer shelves has the following advantages:

[0051] 1. Compared with the existing technology, this application has the characteristics of direct and convenient implementation process, is easy to integrate into the existing automated warehousing system, and has high practical value and application prospects.

[0052] 2. Existing technologies for 3D visual positioning on multi-layer shelves suffer from the disadvantage of increasing cumulative errors. This means that positioning errors accumulate with the number of operations, and initial small errors gradually magnify, ultimately leading to serious problems such as grasping failures or dropped goods. This application effectively reduces error accumulation by updating position information after each grasping operation and comparing it with the previous layer's position information during non-initial positioning. This ensures that positioning results remain stable and accurate even in the complex environment of multi-layer shelves.

[0053] 3. The existing technology lacks a dynamic error correction mechanism, that is, it relies on a single 3D visual positioning result or a pre-set teaching position for verification. Once the storage environment factors or the shelf structure are deformed, the pre-set teaching position will no longer be accurate, causing the positioning result to deviate from the actual position. At the same time, the single positioning verification method cannot dynamically correct the deviation caused by cumulative errors. The present application adopts a dual verification and error correction mechanism. During the first positioning, the 3D positioning result is subtracted from the preset teaching position; during non-first positioning, the 3D positioning result is subtracted from the position information of the previous layer, and when necessary, the difference is calculated again with the teaching position, so that positioning errors can be discovered and corrected in a timely manner and the adverse effects of cumulative errors can be eliminated.

[0054] 4. This application uses a dual verification mechanism to gradually optimize system data by updating location information after each successful positioning, effectively improving the system's adaptability, so that the system can maintain efficient and stable operation in a complex and changing environment. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] The present application will now be further described with reference to the following drawings;

[0056] Figure 1 A schematic diagram of determining the teaching position for applying the 3D visual positioning error correction and alarm method described in this application;

[0057] Figure 2 This is a schematic diagram of positioning error correction for non-first positioning;

[0058] Figure 3 This is a schematic diagram of the 3D visual positioning coordinate system;

[0059] Figure 4 A schematic diagram of a 3D visual positioning error correction alarm system using this application;

[0060] Figure 5 This is a flow chart of the 3D visual positioning error correction and alarm method for multi-layer shelves described in this application. DETAILED DESCRIPTION

[0061] Example 1, as Figure 4 As shown, in order to implement the 3D visual positioning error correction alarm method for multi-layer shelves proposed in this application, this embodiment proposes the following 3D visual positioning error correction alarm system. The system includes:

[0062] The controller includes an image processing unit and a data processing unit. The image processing unit is used to process the images captured by the 3D camera to generate the 3D positioning result coordinates of a certain shelf location. The data processing unit is used to analyze and make decisions based on the positioning comparison results and output action instructions under the premise of implementing a double verification mechanism.

[0063] The robotic arm is used to grab goods from the storage space or place goods in the storage space according to the instructions of the controller; the robotic arm is connected to the controller through a communication interface to receive and execute the action instructions output by the controller;

[0064] A 3D camera, mounted on a robotic arm, is used to obtain 3D point cloud data of a specific location on the shelf;

[0065] An alarm device, including an audible and visual alarm and a control circuit, is used to output an alarm signal including an audible and visual signal when the controller detects that the positioning result exceeds a preset threshold value, so as to prompt the operator to perform inspection and calibration;

[0066] The system database is used to store the manually measured initial positioning data of the first-level shelves and the 3D positioning information before each cargo entry and exit operation. The controller saves each updated position information to the system database as a reference for subsequent data comparison and the execution of actions by the robotic arm.

[0067] like Figures 1 to 5 As shown, the 3D visual positioning error correction and alarm method for multi-layer shelves described in this application includes the following implementation steps:

[0068] Step 1) Determine the teaching position;

[0069] Before the goods are put in and out of the warehouse, the teaching position of the robot arm for each cargo position is marked; specifically,

[0070] Step 1.1, mark the teaching position of each cargo location;

[0071] For the bottom shelf of each row of shelves, the teaching position is manually measured and marked; for the other shelves in the same row, that is, the shelf height data of the other shelves other than the bottom shelf is accumulated layer by layer from the teaching position of the bottom shelf.

[0072] Step 1.2, establish the three-dimensional coordinate system of the shelf;

[0073] Use a 3D camera to take pictures to obtain and establish the three-dimensional coordinate system where the multi-layer shelves are located;

[0074] The current 3D positioning result coordinates (X, Y, Z, W, P, R) of any shelf in the multi-layer shelf can then be identified and calculated for use in each positioning and difference calculation. Among them, X, Y, and Z represent the X-axis coordinates, Y-axis coordinates, and Z-axis coordinates in the three-dimensional coordinate system respectively; W is the angle of rotation around the X-axis, P is the angle of rotation around the Y-axis, and R is the angle of rotation around the Z-axis.

[0075] Step 2), first positioning;

[0076] Before using the robotic arm to perform a loading or unloading operation at a certain cargo location, it takes a photo with a 3D camera to obtain the 3D positioning result coordinates of the current cargo location. The 3D positioning result coordinates are compared with the taught position of the current cargo location. If the difference does not exceed the preset threshold, the robotic arm performs the loading or unloading operation and uses the 3D positioning result coordinates of the current cargo location as the new taught position information and updates it to the system database.

[0077] If the comparison difference exceeds the preset threshold, an alarm signal is triggered and output to prompt the operator to perform manual intervention and calibration; specifically,

[0078] Step 2.1, determine whether it is the first positioning;

[0079] If the current cargo location is the first time to carry out the storage and outgoing operation, it is the first positioning;

[0080] If the current shelf is the bottom shelf on the leftmost side of the entire shelf, it is the first positioning. It should be noted that based on the 3D visual positioning error correction and alarm method for multi-layer shelves proposed in this embodiment, the order of storage and retrieval of shelves on multi-layer shelves is from left to right and from low to high. That is, the storage and retrieval of goods starts from the leftmost and bottom shelf of the multi-layer shelf, and proceeds in sequence.

[0081] Otherwise, it is not the first positioning, and jump directly to step 3);

[0082] Step 2.2, perform positioning information comparison;

[0083] The 3D camera takes a picture to obtain the point cloud data of the current cargo location, and the controller calculates the 3D positioning result coordinates (X, Y, Z, W, P, R) of the current cargo location and the teaching position information (X n ,Y n ,Z n ,W n ,P n ,R n ) to perform difference operations, including

[0084] Difference 1 = |XX n |, difference 2 = |YY n |, difference 3 = |ZZ n |;

[0085] Difference 4 = |WW n |, difference 5 = |PP n |, difference 6 = |RR n |;

[0086] Determine whether any of the above differences 1 to 6 exceeds the preset threshold; if not, the robotic arm performs the goods in and out of the warehouse and updates the 3D positioning result of the current cargo location as the new teaching position information to the system database;

[0087] If so, an alarm signal is triggered and output to prompt the operator to perform manual intervention and calibration;

[0088] Step 3) Non-first positioning;

[0089] First, determine whether the current shelf is on the bottom layer. If so, perform a difference calculation between the 3D positioning result coordinates of the current shelf and the teaching position of the shelf on its left bottom layer stored in the system database. If not, perform a difference calculation between the 3D positioning result coordinates of the current shelf and the teaching position of the shelf on the next layer stored in the system database.

[0090] Secondly, it is determined whether the difference calculation results in any of the above two situations exceed the preset threshold. If not, the robot arm performs the goods in and out of the warehouse and updates the 3D positioning results of the current cargo location as the new teaching position information to the system database. If so, the controller performs a difference calculation between the 3D positioning result coordinates of the current cargo location and the teaching position of the current cargo location stored in the system database.

[0091] Finally, it is determined whether the difference between the 3D positioning result coordinates of the current cargo location and its taught position exceeds the preset threshold; if not, the robot arm performs the goods in and out of the warehouse, and updates the 3D positioning result of the current cargo location as the new taught position information to the system database; if so, it triggers and outputs an alarm signal to prompt the operator to manually intervene and calibrate; specifically,

[0092] Step 3.1, positioning of the current cargo location;

[0093] The controller determines whether the current cargo location is at the bottom layer based on the received current cargo location number;

[0094] If so, perform difference calculation on the 3D positioning result coordinates (X, Y, Z, W, P, R) of the current cargo location and the teaching position (X1, Y1, Z1, W1, P1, R1) of the bottom cargo location on its left stored in the system database, including,

[0095] Difference 1 = |X - X1|, Difference 2 = |Y - Y1|, Difference 3 = |Z - Z1|;

[0096] Difference 4 = |W-W1|, Difference 5 = |P-P1|, Difference 6 = |R-R1|;

[0097] If not, the 3D positioning result coordinates (X, Y, Z, W, P, R) of the current cargo location are combined with the teaching position (X, Y, Z, W, P, R) of the next cargo location stored in the system database. n-1 ,Y n-1 ,Z n-1 ,W n-1 ,P n-1 ,R n-1 ) to perform difference operations, including,

[0098] Difference 1 = |XX n-1 |, difference 2 = |YY n-1 |, difference 3 = |ZZ n-1 |;

[0099] Difference 4 = |WW n-1 |, difference 5 = |PP n-1 |, difference 6 = |RR n-1 |;

[0100] Step 3.2: Determine whether the difference calculation result in any of the above two situations exceeds a preset threshold;

[0101] If any of the difference values ​​1 to 6 in the above two situations do not exceed the preset threshold, the robot arm will perform the goods in and out of the warehouse and update the 3D positioning result (X, Y, Z, W, P, R) of the current cargo location as the new teaching position information to the system database;

[0102] If any of the difference values ​​1 to 6 in the above two situations exceeds the preset threshold, the process proceeds to step 3.3;

[0103] Step 3.3, calculate and judge the difference between the teaching position information;

[0104] Determine the 3D positioning result coordinates (X, Y, Z, W, P, R) of the current cargo location and the teaching position information (X n ,Y n ,Z n ,W n ,Pn ,R n ) performs difference operation; including,

[0105] Difference 1 = |XX n |, difference 2 = |YY n |, difference 3 = |ZZ n |;

[0106] Difference 4 = |WW n |, difference 5 = |PP n |, difference 6 = |RR n |;

[0107] Determine whether any difference calculation result among the above result differences 1 to 6 exceeds a preset threshold;

[0108] If not, the robot arm performs the goods in and out of the warehouse, and updates the 3D positioning result (X, Y, Z, W, P, R) of the current cargo location as the new teaching position information to the system database;

[0109] If so, an alarm signal is triggered and output to prompt the operator to perform manual intervention and calibration.

[0110] This embodiment ensures the accuracy of continuous pick-and-place operations by comparing the current positioning result with the position information of the previous layer, including various scenarios such as first-time and non-first-time, bottom layer and non-bottom layer. If the positioning result does not exceed the threshold, the robot arm continues to perform the pick-and-place operation. If the positioning result exceeds the threshold, the system further compares the current positioning result with the shelf's taught position. If the threshold is exceeded again, an alarm signal is triggered, prompting the operator to intervene.

[0111] If any positioning result exceeds a preset threshold, the system triggers an alarm, emitting audible and visual signals, prompting the operator to perform inspections and calibrations. This alarm is generated based on a dual verification mechanism to ensure that positioning errors are addressed promptly, preventing cumulative errors from leading to system failures.

[0112] After each successful positioning, the system will update the teaching position information of the current cargo location as a reference for the next positioning. By continuously updating and optimizing the position information in the system database, the positioning accuracy and the system's adaptability are improved.

[0113] Through the above-mentioned dual verification mechanism, this application can not only improve the accuracy and stability of 3D visual positioning, but also promptly detect and correct positioning errors, effectively preventing system failures and safety issues caused by cumulative errors.

[0114] The embodiments described above, in conjunction with the accompanying drawings, are merely preferred solutions for achieving the objectives of the present invention. Those skilled in the art will readily be able to derive alternative structures consistent with the design concepts of the present invention based on these insights. Other structural features derived from these alternatives are also intended to fall within the scope of the present invention.

Claims

1. A 3D visual positioning error correction and alarm method for multi-layer shelves, characterized by: The implementation steps include: Step 1) Determine the teaching position; Before the goods are put in or taken out of the warehouse, the teaching position of the robot arm for each cargo location is marked; Step 2), first positioning; Before using the robotic arm to perform a loading and unloading operation at a certain cargo location, it takes a photo with a 3D camera to obtain the 3D positioning result coordinates of the current cargo location. The 3D positioning result coordinates are compared with the taught position of the current cargo location. If the comparison difference does not exceed the preset threshold, the robotic arm performs the loading and unloading operation and uses the 3D positioning result coordinates of the current cargo location as the new taught position information and updates it to the system database. If the comparison difference exceeds the preset threshold, an alarm signal is triggered and output to prompt the operator to manually intervene and calibrate. Step 3) Non-first positioning; First, determine whether the current shelf is on the bottom layer. If so, perform a difference calculation between the 3D positioning result coordinates of the current shelf and the teaching position of the shelf on its left bottom layer stored in the system database. If not, perform a difference calculation between the 3D positioning result coordinates of the current shelf and the teaching position of the shelf on the next layer stored in the system database. Secondly, it is determined whether the difference calculation results in any of the above two situations exceed the preset threshold. If not, the robot arm performs the goods in and out of the warehouse and updates the 3D positioning results of the current cargo location as the new teaching position information to the system database. If so, the controller performs a difference calculation between the 3D positioning result coordinates of the current cargo location and the teaching position of the current cargo location stored in the system database. Finally, it is determined whether the difference between the 3D positioning result coordinates of the current cargo location and its taught position exceeds the preset threshold; if not, the robotic arm performs the goods in and out of the warehouse, and updates the 3D positioning result of the current cargo location as the new taught position information to the system database; if so, an alarm signal is triggered and output to prompt the operator to perform manual intervention and calibration.

2. The 3D visual positioning error correction and alarm method for multi-layer shelves according to claim 1 is characterized in that: The step 1) includes the following steps: Step 1.1, mark the teaching position of each cargo location; For the bottom shelf of each row of shelves, the teaching position is manually measured and marked; for the other shelves in the same row, that is, the shelf height data of the other shelves other than the bottom shelf is accumulated layer by layer from the teaching position of the bottom shelf. Step 1.2, establish the three-dimensional coordinate system of the shelf; Use a 3D camera to take pictures to obtain and establish the three-dimensional coordinate system where the multi-layer shelves are located.

3. The 3D visual positioning error correction and alarm method for multi-layer shelves according to claim 2 is characterized in that: Identify and calculate the current 3D positioning result coordinates (X, Y, Z, W, P, R) of any cargo location, where X, Y, and Z represent the X-axis coordinate, Y-axis coordinate, and Z-axis coordinate in the three-dimensional coordinate system, respectively; W is the angle of rotation around the X-axis, P is the angle of rotation around the Y-axis, and R is the angle of rotation around the Z-axis.

4. The 3D vision positioning error correction and alarm method for multi-layer shelves according to claim 1 is characterized in that: The step 2) includes the following steps: Step 2.1, determine whether it is the first positioning; If the current cargo location is the first time to carry out the storage and outgoing operation, it is the first positioning; If the current location is the bottom shelf on the leftmost side of the entire shelf, it is the first positioning; otherwise, it is not the first positioning and jumps directly to step 3); Step 2.2, perform positioning information comparison; The 3D camera takes a picture to obtain the point cloud data of the current cargo location, and the controller calculates the 3D positioning result coordinates (X, Y, Z, W, P, R) of the current cargo location and the teaching position information (X n ,Y n ,Z n ,W n ,P n ,R n ) performs a difference operation; Difference 1 = |XX n |, difference 2 = |YY n |, difference 3 = |ZZ n |; Difference 4 = |WW n |, difference 5 = |PP n |, difference 6 = |RR n |; Determine whether any of the above difference values ​​1 to 6 exceeds the preset threshold; if not, the robotic arm performs the goods in and out of the warehouse, and updates the 3D positioning result of the current cargo location as the new teaching position information to the system database; if so, trigger and output an alarm signal to prompt the operator to manually intervene and calibrate.

5. The 3D vision positioning error correction and alarm method for multi-layer shelves according to claim 1 is characterized in that: The step 3) includes the following steps: Step 3.1, positioning of the current cargo location; The controller determines whether the current cargo location is at the bottom layer based on the received current cargo location number; If so, perform difference calculation on the 3D positioning result coordinates (X, Y, Z, W, P, R) of the current cargo location and the teaching position (X1, Y1, Z1, W1, P1, R1) of the bottom cargo location on its left stored in the system database; Difference 1 = |X - X1|, Difference 2 = |Y - Y1|, Difference 3 = |Z - Z1|; Difference 4 = |W-W1|, Difference 5 = |P-P1|, Difference 6 = |R-R1|; If not, the 3D positioning result coordinates (X, Y, Z, W, P, R) of the current cargo location are combined with the teaching position (X, Y, Z, W, P, R) of the next cargo location stored in the system database. n-1 ,Y n-1 ,Z n-1 ,W n-1 ,P n-1 ,R n-1 ) performs a difference operation; Difference 1 = |XX n-1 |, difference 2 = |YY n-1 |, difference 3 = |ZZ n-1 |; Difference 4 = |WW n-1 |, difference 5 = |PP n-1 |, difference 6 = |RR n-1 |; Step 3.2: Determine whether the difference calculation result in any of the above two situations exceeds a preset threshold; If any of the difference values ​​1 to 6 in the above two situations do not exceed the preset threshold, the robot arm will perform the goods in and out of the warehouse and update the 3D positioning result (X, Y, Z, W, P, R) of the current cargo location as the new teaching position information to the system database; If any of the difference values ​​1 to 6 in the above two situations exceeds the preset threshold, the process proceeds to step 3.3; Step 3.3: Calculate the difference between the teaching position information and make a judgment; Determine the 3D positioning result coordinates (X, Y, Z, W, P, R) of the current cargo location and the teaching position information (X n ,Y n ,Z n ,W n ,P n ,R n ) performs a difference operation; Difference 1 = |XX n |, difference 2 = |YY n |, difference 3 = |ZZ n |; Difference 4 = |WW n |, difference 5 = |PP n |, difference 6 = |RR n |; Determine whether any difference calculation result among the above result differences 1 to 6 exceeds a preset threshold; If not, the robot arm performs the goods in and out of the warehouse, and updates the 3D positioning result (X, Y, Z, W, P, R) of the current cargo location as the new teaching position information to the system database; If so, an alarm signal is triggered and output to prompt the operator to perform manual intervention and calibration.

Images