Automatic collection and verification method for warehouse-in and warehouse-out information of regularly-packaged goods

Abstract

The invention relates to the technical field of stock or inventory management, and discloses an automatic collection and verification method for warehouse-in and warehouse-out information of regularly-packaged goods, and the method comprises the steps: firstly, carrying out the centralized utilization of manpower in a warehouse-in and carrying process, placing the goods to a position and a posture suitable for code scanning and size measurement under the assistance of a posture guider when the goods are carried, and carrying out the verification of the warehouse-in and warehouse-out information of the goods; the size of the goods is measured and compared with a standard value in a set form, and then the code scanning information is combined, so that whether the goods in storage are required goods or not can be accurately judged, and therefore, a size reference is provided for the subsequent clamping process of the mechanical arm, and storage errors are avoided; the goods cannot be turned over due to the fact that the degree of freedom of the mechanical arm is small during warehouse-out, the goods can be kept in the warehouse-in posture after the goods are put back to the detection channel and make contact with the posture guider, and therefore code scanning and size measuring can be conducted again, and whether the warehouse-out goods are needed goods or not can be accurately judged. The above points are combined, so that the collection and verification process of scattered and error-prone warehouse-in and warehouse-out information is automatic, and input support is provided for full-automatic warehouse-in and warehouse-out.

Description

Technical Field

[0001] The present invention relates to the technical field of inventory or stock management, and particularly to an automatic acquisition and verification method for the inbound and outbound information of regular-packaged goods. Background Art

[0002] With the development of automation technology, more and more inbound and outbound processes can be automated, thus greatly improving production efficiency and saving labor. However, there are still some inbound and outbound processes that require a large number of manual operations. The most typical case is the pharmacy in a hospital. In theory, the outbound process (i.e., dispensing drugs) in this scenario can be fully automated because the demand is very certain. Each patient only needs to pick up the drugs prescribed by the doctor for himself. However, in practice, the drug dispensing process cannot be automated. A large number of manual drug dispensing windows still need to be opened, with a pharmacist assigned to each window, and then patients have to queue up for a long time to get their drugs. This process is not only painful for patients, but also wastes a lot of human resources by having a large number of pharmacists engaged in such simple and repetitive work.

[0003] Modern servo systems and PID control systems have been able to pick up objects accurately and quickly at a fixed point. Therefore, there is no difficulty in the physical layer of this process. The bottleneck restricting this process lies in: it is impossible to simply and quickly identify goods by machines.

[0004] When people identify goods, they adopt a series of complex and mutually verifiable methods, including but not limited to observing the text and appearance on them, weighing their weight, estimating their size, smelling their odor, etc. This process requires the use of a machine tool with dozens of degrees of freedom like a human hand, multiple sets of complex and delicate sensors throughout the body, and multiple sets of neural network models pre-trained in the human brain for decades (including various different models in specific fields and general fields). In terms of current technology, implementing this process with machines is not only extremely costly but also far less efficient than humans. For example, machine vision can identify specific objects such as human faces, but a machine for face recognition cannot identify dogs. To identify dogs, a completely different set of machine vision devices is required, and this device will also misidentify cats as dogs. In warehouse management, there are thousands of types of goods to be identified, and these goods are inanimate objects and will not actively align the side containing their features with the camera like face recognition.

[0005] Of course, there are also some identification - assisting marks on the goods, such as product barcodes and QR codes. However, the inventor found that in many cases, these marks are not unique, which leads to incorrect identification and out - of - warehouse operations. Taking product barcodes as an example, theoretically, each product should have a unique barcode. Even for the same product, changing the specification will result in different barcodes. Even a box of milk will have a different barcode from a carton of milk. But in actual operations, there are various situations that can cause barcode duplication. In real life, in bundled sales and promotional activities, it is often the case that multiple products are packed together for sale (a typical example is the so - called "one piece" of beer). At this time, the barcode of the multiple products bundled together is the barcode of a single product (of course, some large supermarkets have the ability to re - spray codes, but most still rely on manual identification). This situation is even more common in pharmacies, where many drugs are packaged in a bundle with plastic film / tape. Humans can easily identify this problem, but machines cannot. In daily life, only express deliveries have unique identification marks, so automation has been achieved very early. In addition, automation has only been achieved for some small - packaged beverages and foods, and the impact of misidentifying these goods is not significant.

[0006] Based on the above findings, if it is possible to identify the size of the goods while identifying barcodes and QR codes (theoretically, weighing could also be used, but in actual operations, it is found that weighing is not feasible because there are various factors affecting the result, such as an indefinite amount of filling materials), then the goods can be correctly identified. However, since the robotic arm does not have binocular vision and tactile sense like a human, and its degree of freedom is also insufficient, it can transfer goods but cannot assist in measurement (it is very difficult to adjust different goods to specific poses). Summary of the Invention

[0007] The present invention provides an automatic acquisition and verification method for the inbound and outbound information of regularly - packaged goods.

[0008] The technical problem to be solved is that due to the inability to rely on machines to simply and quickly identify goods, and the fact that the same barcode / QR code may correspond to different goods, the inbound and outbound of goods cannot be fully automated.

[0009] To solve the above - mentioned technical problem, the present invention adopts the following technical solution: An automatic acquisition and verification method for the inbound and outbound information of regularly - packaged goods, which is used to determine whether the goods are the required goods when the regularly - packaged goods are inbound and outbound. The regularly - packaged goods are goods with a cuboid package and have a scannable barcode and / or QR code on the outer surface. The barcode and the QR code are uniformly recorded as information codes, and the information codes contain information including the name of the regularly - packaged goods. The automatic acquisition and verification method includes the following steps: Step 1: When warehousing is required, release the warehousing demand, specifying the name and quantity of the regular-packaged goods to be warehoused; count the information codes and the length, width, and height dimensions of the goods to be warehoused, establish a database, and then input the database into the server. Step 2: Set up a detection channel. One end of the detection channel is close to the robotic arm and the shelf, and the other end is close to the operator and is provided with an automatic door. The automatic door is a door that can report its opening and closing status to the server and is controlled by the server for opening and closing; the detection channel is equipped with three laser range sensors for measuring the length, width, and height data of the regular-packaged goods, and one or more barcode scanning cameras for scanning barcodes. The laser range sensors, barcode scanning cameras, and automatic door are all electrically connected to the server. The lasers emitted by the three laser range sensors are perpendicular to each other, and a pose guide for controlling the pose of the regular-packaged goods is also set in the detection channel. Step 3: Denote the three mutually perpendicular faces of the regular-packaged goods to be warehoused as detection faces. The on-site operator opens the automatic door and places a regular-packaged goods to be warehoused into the detection channel, using the pose guide to make the three detection faces of the regular-packaged goods perpendicular to the light beams of the adjacent laser range sensors respectively, and making the information code within the shooting range of a barcode scanning camera. Use the three laser range sensors to measure the length, width, and height data of the regular-packaged goods in the detection channel respectively, and use the barcode scanning camera to scan the information code. Step 4: The server determines whether the information code of the regular-packaged goods in the detection channel matches the information code of the replenishment demand. If it matches, then determine whether the length, width, and height values match. If they match, the robotic arm will place the regular-packaged goods on the shelf and record the shelf position; then close the automatic door of the detection channel and record one warehousing behavior. Denote the face of the regular-packaged goods facing the automatic door as the positioning face. During the process of transferring the regular-packaged goods from the detection channel to the shelf, the bottom surface should always face downwards, and after the regular-packaged goods are transferred to the shelf, the positioning face faces outside the shelf. Step 5: Repeat Steps 2 to 4 until the warehousing is completed. Step 6: When the goods need to be shipped out and all the regular-packaged goods to be shipped out are already on the shelf, release the shipping demand, specifying the name and quantity of the regular-packaged goods to be shipped out. Step 7: The robotic arm takes out a regularly packaged item from the position recorded in Step 5 and places it in the inspection channel. The server determines whether the information code of the regularly packaged item in the inspection channel matches the information code of the regularly packaged item in the outbound requirement. If they match, it further determines whether the length, width, and height values match. If they match, the automatic door is opened. After the operator takes out the regularly packaged item, the automatic door is closed and the server records an outbound behavior (generally, an automatic door closing instruction is sent 2 seconds after the laser detects no reflection from the goods). If they do not match, the robotic arm puts the regularly packaged item back on the shelf and records an abnormal re-taking and placing behavior at this location for the administrator to review and repair later; During the process of transferring the regularly packaged item from the shelf to the inspection channel and from the inspection channel to the shelf, the bottom surface should always face downwards. After the regularly packaged item is transferred to the shelf, the positioning surface faces outside the shelf. After the regularly packaged item is transferred to the inspection channel, the positioning surface faces the automatic door; Step 8: Repeat Step 7 until the outbound process is completed.

[0010] Furthermore, the pose guide is an L-shaped hem set at the inner bottom of the inspection channel. One side of the L-shaped hem is parallel to the automatic door, close to the automatic door, and is denoted as the depth control side, and the other side is close to one side wall of the inspection channel and is denoted as the offset control side; The three laser range sensors are respectively denoted as the top sensor, the side sensor, and the depth sensor; The top sensor is set at the top of the inspection channel and the laser beam irradiates the regularly packaged item downward. The distance between the top sensor and the inner surface of the opposite side of the inspection channel is denoted as Z0, and the distance between the top sensor and the regularly packaged item is denoted as Z1, and Z = Z0 - Z1; The side sensor is set on the side wall of the inspection channel away from the offset control side, and the laser beam of the side sensor irradiates the regularly packaged item horizontally parallel to the automatic door. The distance between the side sensor and the inner surface of the opposite side of the inspection channel is denoted as X0, and the distance between the side sensor and the regularly packaged item is denoted as X1, and X = X0 - X1; The depth sensor is located at one end of the inspection channel close to the shelf. The depth sensor is set outside the inspection channel and avoids the working range of the robotic arm. The laser beam of the depth sensor irradiates the regularly packaged item horizontally perpendicular to the automatic door. The distance between the depth sensor and the inner surface of the opposite side of the inspection channel is denoted as Y0, and the distance between the depth sensor and the regularly packaged item is denoted as Y1, and Y = Y0 - Y1.

[0011] Furthermore, each inspection channel is equipped with three barcode scanning cameras. The three barcode scanning cameras are respectively set close to the three laser range sensors, and the lens orientation is the same as that of the adjacent laser range sensor; In step 4, the on-site operator places the neatly packaged goods into the inspection channel so that the neatly packaged goods are close to the depth control edge and the offset control edge at the same time, and the side of the neatly packaged goods with the information code is irradiated by a laser beam.

[0012] Furthermore, the three values of length, width and height of the neatly packaged goods in the database corresponding to the detection channel are recorded as x, y and z respectively; the unordered set composed of x, y and z is recorded as A, and the unordered set composed of X, Y and Z is recorded as B; Step 4: When comparing the length, width and height values, if A=B, then the length, width and height values are considered to match; Step 7: When comparing the length, width and height values, if Z∈A, an element m that is not Z and not greater than X can be found in set A, and an element n that is not Z and not n and not greater than Y can be found in set A, then the length, width and height values are considered to match.

[0013] Furthermore, in step 4, when comparing the length, width and height values, sort the elements in set A from large to small or from small to large, and then sort the elements in set B in the same way, and then compare the sizes of the elements in set A and set B one by one in order. If the sizes of all elements in the two sets are the same, then the length, width and height values are considered to match; Step 7: When comparing the length, width and height values, Z is compared with the elements in set A one by one. If the value of Z is consistent with one of the elements, the elements in set A except Z are recorded as a and b; if X is not less than a and Y is not less than b, or X is not less than b and Y is not less than a, then the length, width and height values are considered to match.

[0014] Furthermore, the robot arm grasps the neatly packaged goods through the gripper, a point is selected between the claws of the gripper as the center of the gripper, the distance between the center of the gripper and each claw is the same, the robot arm is provided with a positioning device for real-time positioning of the coordinates of the center of the gripper, the coordinates of the intersection of the depth control edge and the offset control edge are marked as (0, 0, 0), and the coordinates of the geometric center of the neatly packaged goods are marked as (X / 2, Y / 2, Z / 2); When clamping the neatly packaged goods, the opening width of the clamping jaws is greater than X, and the center of the clamping jaws is located between the geometric center of the neatly packaged goods and the positioning surface and is lower than the geometric center of the neatly packaged goods; In step 4, when the gripper just grips and picks up the neatly packaged goods, the center coordinates of the gripper at this time (x1, y1, z1) are recorded; In step seven, when the neatly packaged goods are placed into the inspection channel, if the center of the gripper reaches (x1, y1, z1), the position and posture of the neatly packaged goods are considered qualified, and the gripper can be released.

[0015] Furthermore, the automatic collection and verification method uses multiple detection channels and multiple sets of robotic arms, so that the warehousing and the outbound delivery of neatly packaged goods stored in the shelves can be carried out simultaneously.

[0016] Compared with the prior art, the automatic acquisition and verification method for the information of regular-packaged goods in and out of the warehouse of the present invention has the following beneficial effects: In the present invention, the manpower during the inbound handling process is first centralized. When transporting the goods to the detection channel, the goods are placed in a position and posture suitable for scanning and measuring the size with the assistance of the pose guide, and the size of the goods is measured and compared with the standard value in the form of a set (because it is not known which measured size corresponds to which size in the standard value). Then, combined with the scanned code information, it is possible to accurately determine whether the inbound goods are the required goods, thereby providing a size reference for the subsequent gripping process of the robotic arm and avoiding inbound errors. When outbound, since the degree of freedom of the robotic arm is very small and the goods cannot be flipped, when the goods are placed back in the detection channel and come into contact with the pose guide, the goods will maintain the pose during inbound, so that the goods can be scanned and measured again to accurately determine whether the outbound goods are the required goods.

[0017] Combining the above points, the scattered and labor-consuming processes of outbound, inbound and outbound information acquisition and verification are automated. Only a small amount of manpower is required for the handling work during inbound (only one person needs to put the goods into the detection channel and align them properly), thus greatly improving the efficiency of the inbound and outbound processes and realizing automatic inbound and outbound. At the same time, this technology can be seamlessly integrated with existing systems such as inventory management software and POS systems, without the need for relevant software development for this technology, improving the operation efficiency and data visibility.

[0018] At the same time, since the entire process is compatible with different goods, the equipment using this method can be used for multiple purposes, such as being used as a courier cabinet, a vending machine, and an automatic pharmacy at the same time. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 It is a schematic diagram of the automatic acquisition and verification method for the information of regular-packaged goods in and out of the warehouse of the present invention, and the perspective in the figure is a top view; Figure 2 It is a schematic structure of the detection channel Figure 1 ; Figure 3 It is a schematic structure of the detection channel Figure 2 ; In the figure, 1 - detection channel, 2 - regular-packaged goods, 3 - automatic door, 4 - laser range finder, 5 - barcode scanning camera, 6 - pose guide, 7 - robotic arm. DETAILED DESCRIPTION OF THE INVENTION

[0020] As Figures 1-3As shown in the figure, an automatic acquisition and verification method for the inbound and outbound information of regularly packaged goods is used to determine whether the regularly packaged goods 2 are the required goods when they are inbound and outbound. The regularly packaged goods 2 are goods with a cuboid package and have a scannable barcode and / or QR code on the outer surface. The barcode and QR code are collectively referred to as information codes, and the information codes contain information including the name of the regularly packaged goods 2. The automatic acquisition and verification method includes the following steps: Generally, the information code includes information such as the product name, origin, and specifications. However, the origin is useless for inbound and outbound, and the specifications may change due to offline operations (such as packing multiple items or removing a part). Therefore, the product name is used in this embodiment, and the specifications are measured and compared on-site. As for the shape of the package, it is difficult to measure and compare irregularly shaped packages. Therefore, the most common cuboid (including cube, allowing rounded corners and chamfers) package is selected for measurement here, so only the length, width, and height need to be measured.

[0021] Step 1: When inbound is required, release an inbound demand, specifying the name and quantity of the regularly packaged goods 2 to be inbound; count the information codes and the length, width, and height dimensions of the goods to be inbound, establish a database, and then enter the database into the server; Here, the server can be a single-chip microcomputer or one or more computers. When inbound, first determine the demand, then count according to the demand, and enter it into the server. It is also best to have a touch screen electrically connected to the server on-site to adjust the demand, open and close the automatic door 3, etc. on-site.

[0022] Step 2: Set up a detection channel 1. One end of the detection channel 1 is close to the robotic arm 7 and the shelf, and the other end is close to the operator and is provided with an automatic door 3. The automatic door 3 is a door that can report its opening and closing status to the server and is controlled by the server to open and close; the detection channel 1 is equipped with three laser range sensors 4 for measuring the length, width, and height data of the regularly packaged goods 2, and one or more barcode scanning cameras 5 for scanning codes. The laser range sensors 4, the barcode scanning cameras 5, and the automatic door 3 are all electrically connected to the server; Here, the operator refers to the porter. The porter puts the goods into the porter and positions them correctly, and then they are automatically stored in the warehouse and the statistical verification information is obtained; The lasers emitted by the three laser range sensors 4 are perpendicular to each other, and a pose guide 6 for controlling the pose of the regularly packaged goods 2 is also set in the detection channel 1; In this way, the measured results form a set. No matter how the regularly packaged goods 2 are placed, the set formed by the measured length, width, and height data will not change.

[0023] In addition, the automatic door 3 should have the function of preventing accidental injury, and monitor in real time whether there are foreign objects in the cargo opening to avoid accidental injury to humans and damage to cargo during operation. At the same time, the automatic door 3 should be equipped with a water diversion groove with diversion and waterproof functions to ensure that the functions of the information collection and verification area are not affected by water ingress.

[0024] Step 3: The three mutually perpendicular surfaces of the neatly packaged goods 2 to be stored are recorded as detection surfaces. The on-site operator opens the automatic door 3 and puts a neatly packaged good 2 to be stored into the detection channel 1. The posture guide 6 is used to make the three detection surfaces of the neatly packaged goods 2 perpendicular to the light beams of the adjacent laser distance measuring sensors 4, and the information code is located within the shooting range of a code scanning camera 5. The three laser distance measuring sensors 4 are used to measure the length, width and height data of the neatly packaged goods 2 in the detection channel 1 respectively, and the barcode scanning camera 5 is used to scan the information code; The measurement results here do not allow us to know which dimension is the length, which dimension is the width, and which dimension is the height, because under the premise of being able to scan the code and measure the size, there are many different ways to place the neatly packaged goods 2.

[0025] Step 4: The server determines whether the information code of the neatly packaged goods 2 in the detection channel 1 matches the information code of the neatly packaged goods 2 in the replenishment demand. If so, it determines whether the length, width and height values match. If so, the robot arm 7 puts the neatly packaged goods 2 on the shelf and records the shelf position; then the automatic door 3 of the detection channel 1 is closed and a warehousing behavior is recorded; When recording, it should be linked with the inventory management software to update the inventory information synchronously. When entering the warehouse, find an empty space based on the inventory management software, and when leaving the warehouse, let the robot arm 7 find the closest to the detection channel 1 from multiple neatly packaged goods 2 that meet the conditions.

[0026] The surface of the neatly packaged goods 2 facing the automatic door 3 is recorded as the positioning surface. During the process of transferring the neatly packaged goods 2 from the inspection channel 1 to the shelf, the bottom surface must always be downward, and after the neatly packaged goods 2 are transferred to the shelf, the positioning surface faces outside the shelf; Here, the bottom surface and the positioning surface are used to limit the neatly packaged goods 2 from flipping during the transportation process. They are put into the shelf in the same way as they are taken out from the detection channel 1, and then returned to the detection channel 1 in the same posture. If they are flipped, not only will the measurement results lose their meaning, but the information code may not be scanned. The existing robotic arm 7 has very few degrees of freedom and usually does not have the ability to flip the neatly packaged goods 2. Even if they have it, there is no place where they need to be flipped during the transportation process.

[0027] Step 5: Repeat steps 2 to 4 until the storage is completed; Step 6: When all the regular-packed goods 2 that need to be shipped out are located on the shelf, issue a shipping requirement, specifying the name and quantity of the regular-packed goods 2 that need to be shipped out. Step 7: The robotic arm 7 takes out a regular-packed good 2 from the position recorded in Step 5 and places it in the inspection channel 1. The server determines whether the information code of the regular-packed good 2 in the inspection channel 1 matches the information code of the regular-packed good 2 in the shipping requirement. If they match, it further determines whether the length, width, and height values match. If they match, the automatic door 3 is opened. After the operator takes out the regular-packed good 2, the automatic door 3 is closed and the server records a shipping behavior. If they do not match, the robotic arm 7 returns the regular-packed good 2 to the shelf and records an abnormal repositioning behavior at this storage location for the administrator to review and repair later. During the process of transferring the regular-packed good 2 from the shelf to the inspection channel 1 and from the inspection channel 1 to the shelf, the bottom surface should always face downwards. After the regular-packed good 2 is transferred to the shelf, the positioning surface faces outwards from the shelf. After the regular-packed good 2 is transferred to the inspection channel 1, the positioning surface faces the automatic door 3. The operator in this step refers to the user of the regular-packed good 2. Shipping means the process of the regular-packed good 2 reaching the user's hands from the shelf. The so-called "issuing a shipping requirement" is to input what one wants on the touch screen to the server, and the server notifies the robotic arm 7 to pick up goods one by one for the user. When picking up goods, if it is an express delivery, the goods are single and not substitutable. If the picked-up goods fail the inspection, there is no chance for replacement. A grid shelf should be used, with one express delivery placed in each grid. If it fails to be recognized multiple times, it is returned to its original position and the staff is contacted for inspection and repair. If the regular-packed good 2 is not a unique item, multiple of the same kind of goods can be placed in each grid, and if it fails to be recognized, another one is selected.

[0028] Step 8: Repeat Step 7 until the shipping is completed.

[0029] The pose guide 6 is an L-shaped hem provided at the inner bottom of the inspection channel 1. One side of the L-shaped hem is parallel to the automatic door 3, is close to the automatic door 3, and is denoted as the depth control side, and the other side is close to one side wall of the inspection channel 1 and is denoted as the offset control side. This structure can assist the human hand and the robotic arm 7 in adjusting the position and pose of the regular-packed good 2 without jamming the robotic arm 7.

[0030] The three laser range sensors 4 are respectively denoted as the top sensor, the side sensor, and the depth sensor. The top sensor is provided at the top of the inspection channel 1 and the laser beam irradiates the regular-packed good 2 downward. The distance between the top sensor and the inner surface of the opposite inspection channel 1 is denoted as Z0, the distance between the top sensor and the regular-packed good 2 is denoted as Z1, and Z = Z0 - Z1. The side sensor is arranged on the side wall of the detection channel 1 away from the offset control edge. The laser beam of the side sensor is parallel to the automatic door 3 and horizontally irradiates the regular packaged goods 2; the distance between the side sensor and the inner surface of the opposite detection channel 1 is denoted as X0, the distance between the side sensor and the regular packaged goods 2 is denoted as X1, and X = X0 - X1; The depth sensor is located at one end of the detection channel 1 close to the shelf. The depth sensor is arranged outside the detection channel 1 and avoids the working range of the robotic arm 7. The laser beam of the depth sensor is perpendicular to the automatic door 3 and horizontally irradiates the regular packaged goods 2; the distance between the depth sensor and the inner surface of the opposite detection channel 1 is denoted as Y0, the distance between the depth sensor and the regular packaged goods 2 is denoted as Y1, and Y = Y0 - Y1.

[0031] Note that generally when Z1 < Z0, X1 < X0, and Y1 < Y0 are satisfied simultaneously and remain stable for 2 seconds without change (that is, 2 seconds after placing the regular packaged goods 2 and aligning it), the measurement is considered completed.

[0032] If the regular packaged goods 2 is too small to block all three laser beams, the measurement cannot be completed and it cannot be stored in the warehouse. In this case, even if it is stored in the warehouse, the regular packaged goods 2 is too small to be picked up by the robotic arm 7. It indicates that there is a problem with the warehousing requirement, and the warehousing requirement should be changed on-site and this regular packaged goods 2 should be deleted. If there is really no other way, its packaging should be adjusted later.

[0033] Each detection channel 1 is equipped with three barcode scanning cameras 5. The three barcode scanning cameras 5 are respectively arranged close to the three laser distance sensors 4, and the lens orientation is the same as that of the adjacent laser distance sensor 4; This setting method greatly simplifies the workload of the porter to align the regular packaged goods 2. As long as one side of the regular packaged goods 2 with the information code is irradiated by a laser beam, it can ensure that it can be scanned.

[0034] In step four, the on-site operator places the regular packaged goods 2 into the detection channel 1, makes the regular packaged goods 2 closely adhere to both the depth control edge and the offset control edge, and makes the side of the regular packaged goods 2 with the information code be irradiated by a laser beam.

[0035] The three values of the length, width, and height of the regular packaged goods 2 corresponding to the detection channel 1 in the database are respectively denoted as x, y, and z; the unordered set composed of x, y, and z is denoted as A, and the unordered set composed of X, Y, and Z is denoted as B; In step four, when comparing the length, width, and height values, if A = B, it is considered that the length, width, and height values match; Here, it is not known which of X, Y, and Z is the length / width / height, so it is impossible to compare each value correspondingly. However, the unordered set composed of these data will not change with the placement method. Therefore, this method can be used to reduce the requirements during placement while meeting the verification requirements.

[0036] Step 7 When comparing the length, width, and height values, if Z ∈ A, an element m that is non-Z and not greater than X can be found in set A, and an element n that is non-Z, non-n, and not greater than Y can be found in set A, then the length, width, and height values are considered to match.

[0037] The robotic arm 7 is different from a human hand. It has no tactile sense and cannot ensure that the regular packaged goods 2 fit perfectly with the pose guide 6. However, the measurement structure and the pose guide 6 in this embodiment make the measured Z still reliable in this case. If there are deviations in X and Y, they must be positive deviations. Therefore, in this case, if the size in one direction is consistent with the standard value recorded in the database, and the sizes in the other two directions are larger, the correct size can still be ensured.

[0038] Because the regular packaged goods 2 are taken from the shelf position recorded by the server during outbound. The regular packaged goods 2 at this position have been verified for size during inbound. For the regular packaged goods 2 that meet the above conditions during outbound, they are either the correct goods or new goods formed by repackaging multiple correct goods. However, the latter situation cannot exist at the picking position.

[0039] What is actually mainly avoided here is the situation where the packaging box is crushed and only a part of the packaging box with the information code position is clamped over.

[0040] In actual operation, due to the comparison of unordered sets, it is not easy to implement for some programming languages. Therefore, the following equivalent algorithm can be adopted here: Step 4 When comparing the length, width, and height values, sort the elements in set A from largest to smallest or from smallest to largest, then sort the elements in set B in the same sorting way, and then compare the element sizes in set A and set B one by one in order. If the sizes of all elements in the two sets are the same, then the length, width, and height values are considered to match; When making comparisons here, a certain deviation is allowed. For example, within a deviation range of ±3 mm, they are all considered to be of the same size, and the deviation can be ignored. This deviation is caused by unavoidable factors such as measurement errors. Therefore, the specific allowable range is determined by the factors that may affect the measurement accuracy on site.

[0041] Step 7 When comparing the length, width, and height values, compare Z with each element in set A one by one. If the value is the same as one of the elements, then record the elements in set A other than Z as a and b; if X is not less than a and Y is not less than b, or X is not less than b and Y is not less than a, then the length, width, and height values are considered to match.

[0042] Here, X is not less than a and Y is not less than b, or X is not less than b and Y is not less than a, that is, the values of X and Y are allowed to have non-negligible positive deviations, and the positive deviations are still controlled within a certain range, such as within two centimeters. The specific allowable range is determined by the operating accuracy of the robot arm 7.

[0043] The robot arm 7 grabs the neatly packaged goods 2 through the gripper, and a point is selected between the claws of the gripper as the center of the gripper. The distance between the center of the gripper and each claw is the same. The robot arm 7 is equipped with a positioning device for real-time positioning of the coordinates of the center of the gripper. The coordinates of the intersection of the depth control edge and the offset control edge are marked as (0, 0, 0), and the coordinates of the geometric center of the neatly packaged goods 2 are marked as (X / 2, Y / 2, Z / 2); The positioning device can adopt a variety of structures such as three mutually perpendicular magnetic scales, grating scales, laser rangefinder arrays, etc., and can be selected according to site needs in actual operation.

[0044] When clamping the neatly packaged goods 2, the opening width of the clamping jaws is greater than X, and the center of the clamping jaws is located between the geometric center of the neatly packaged goods 2 and the positioning surface and is lower than the geometric center of the neatly packaged goods 2; In step 4, when the gripper just grips and grabs the neatly packaged product 2, the center coordinates of the gripper at this time (x1, y1, z1) are recorded; In step seven, when the neatly packaged goods 2 are placed in the inspection channel 1, if the center of the clamp reaches (x1, y1, z1), the position and posture of the neatly packaged goods 2 are considered qualified, and the clamp can be released.

[0045] Here, coordinates are used to assist the clamping and movement of the robot arm 7, so that the position can be more accurate and the clamping is not easy to fall off.

[0046] The automatic collection and verification method uses multiple detection channels 1 and multiple sets of robotic arms 7, so that the storage and the outbound delivery of neatly packaged goods 2 stored in the shelves can be carried out simultaneously.

[0047] For example, aspirin can be replenished in the warehouse while patients can take amoxicillin out. Or a package can be put in while a package can be taken out.

[0048] The embodiments described above are merely descriptions of preferred implementation modes of the present invention and are not intended to limit the scope of the present invention. Without departing from the design spirit of the present invention, various modifications and improvements made to the technical solutions of the present invention by ordinary technicians in this field should all fall within the protection scope determined by the claims of the present invention.

Claims

1. An automatic acquisition and verification method for the information of regular-packaged goods in and out of the warehouse, which is used to judge whether the regular-packaged goods (2) are the required goods when they are warehoused and out of the warehouse. The regular-packaged goods (2) are goods with a cuboid package, and have a scannable barcode and / or two-dimensional code on the outer surface. The barcode and the two-dimensional code are collectively recorded as information codes, and the information codes carry information including the name of the regular-packaged goods (2). It is characterized in that: The automatic acquisition and verification method includes the following steps: Step 1: When warehousing is required, release the warehousing requirement, specifying the name and quantity of the regular-packaged goods (2) to be warehoused; count the information codes and the length, width, and height dimensions of the goods to be warehoused, establish a database, and then input the database into the server; Step 2: Set up a detection channel (1). One end of the detection channel (1) is close to the robotic arm (7) and the shelf, and the other end is close to the operator and is provided with an automatic door (3). The automatic door (3) is a door that can report its opening and closing status to the server and is controlled by the server for opening and closing; the detection channel (1) is equipped with three laser range sensors (4) for measuring the length, width, and height data of the regular-packaged goods (2), and one or more barcode scanning cameras (5) for scanning barcodes. The laser range sensors (4), barcode scanning cameras (5), and the automatic door (3) are all electrically connected to the server; The lasers emitted by the three laser range sensors (4) are perpendicular to each other, and a pose guide (6) for controlling the pose of the regular-packaged goods (2) is also set in the detection channel (1); Step 3: The three mutually perpendicular faces of the regular-packaged goods (2) to be warehoused are denoted as detection faces. The on-site operator opens the automatic door (3) and places a regular-packaged goods (2) to be warehoused into the detection channel (1). Use the pose guide (6) to make the three detection faces of the regular-packaged goods (2) perpendicular to the light beams of the adjacent laser range sensors (4) respectively, and make the information code within the shooting range of one barcode scanning camera (5); Use the three laser range sensors (4) to measure the length, width, and height data of the regular-packaged goods (2) in the detection channel (1) respectively, and use the barcode scanning camera (5) to scan the information code; Step 4: The server determines whether the information code of the regular-packaged goods (2) in the detection channel (1) matches the information code of the regular-packaged goods (2) in the replenishment requirement. If it matches, then determine whether the length, width, and height values match. If they match, the robotic arm (7) places the regular-packaged goods (2) on the shelf and records the shelf position; then close the automatic door (3) of the detection channel (1) and record one warehousing behavior; The face of the regular-packaged goods (2) facing the automatic door (3) is denoted as the positioning face. During the process of transferring the regular-packaged goods (2) from the detection channel (1) to the shelf, the bottom surface should always face downwards, and after the regular-packaged goods (2) is transferred to the shelf, the positioning face faces outside the shelf; Step 5: Repeat steps two to four until the warehousing is completed; Step 6: When the goods need to be shipped out and all the regular-packaged goods (2) to be shipped out are already on the shelf, release the shipping requirement, specifying the name and quantity of the regular-packaged goods (2) to be shipped out; Step 7: The robotic arm (7) takes out a regular packaged product (2) from the position recorded in Step 5 and places it in the inspection channel (1). The server determines whether the information code of the regular packaged product (2) in the inspection channel (1) matches the information code of the regular packaged product (2) in the outbound requirement. If they match, it further determines whether the length, width, and height values match. If they match, the automatic door (3) is opened. After the operator takes out the regular packaged product (2), the automatic door (3) is closed and the server records an outbound behavior. If they do not match, the robotic arm (7) puts the regular packaged product (2) back on the shelf and records an abnormal pick-and-place behavior at this storage location for the administrator to review and repair afterwards. During the transfer of the regular packaged product (2) from the shelf to the inspection channel (1) and from the inspection channel (1) to the shelf, the bottom surface should always face downwards. After the regular packaged product (2) is transferred to the shelf, the positioning surface faces outside the shelf. After the regular packaged product (2) is transferred to the inspection channel (1), the positioning surface faces the automatic door (3). Step 8: Repeat Step 7 until the outbound process is completed.

2. The automatic acquisition and verification method for the inbound and outbound information of regular-packaged goods according to claim 1, wherein: The pose guide (6) is an L-shaped hem provided at the inner bottom of the inspection channel (1). One side of the L-shaped hem is parallel to the automatic door (3), close to the automatic door (3), and is denoted as the depth control side, and the other side is close to one side wall of the inspection channel (1) and is denoted as the offset control side. The three laser range sensors (4) are respectively denoted as the top sensor, the side sensor, and the depth sensor. The top sensor is provided at the top of the inspection channel (1) and the laser beam irradiates the regular packaged product (2) downward. The distance between the top sensor and the inner surface of the opposite inspection channel (1) is denoted as Z0, and the distance between the top sensor and the regular packaged product (2) is denoted as Z1. Z = Z0 - Z1. The side sensor is provided on the side wall of the inspection channel (1) away from the offset control side. The laser beam of the side sensor irradiates the regular packaged product (2) horizontally parallel to the automatic door (3). The distance between the side sensor and the inner surface of the opposite inspection channel (1) is denoted as X0, and the distance between the side sensor and the regular packaged product (2) is denoted as X1. X = X0 - X1. The depth sensor is located at one end of the inspection channel (1) close to the shelf. The depth sensor is provided outside the inspection channel (1) and avoids the working range of the robotic arm (7). The laser beam of the depth sensor irradiates the regular packaged product (2) horizontally perpendicular to the automatic door (3). The distance between the depth sensor and the inner surface of the opposite inspection channel (1) is denoted as Y0, and the distance between the depth sensor and the regular packaged product (2) is denoted as Y1. Y = Y0 - Y1.

3. A method for automatically collecting and verifying the inbound and outbound information of regular packaged goods according to claim 2, characterized in that: Each inspection channel (1) is equipped with three barcode scanning cameras (5). The three barcode scanning cameras (5) are respectively arranged close to the three laser range sensors (4), and the lens orientation is the same as that of the adjacent laser range sensor (4). In Step 4, the on-site operator places the regular packaged product (2) into the inspection channel (1) so that the regular packaged product (2) is in contact with both the depth control side and the offset control side at the same time, and the side with the information code of the regular packaged product (2) is irradiated by a beam of laser.

4. The automatic acquisition and verification method for the inbound and outbound information of regular packaged goods according to claim 2, wherein: The three values of the length, width, and height of the regular packaged goods (2) corresponding to the detection channel (1) in the database are respectively denoted as x, y, and z; the unordered set composed of x, y, and z is denoted as A, and the unordered set composed of X, Y, and Z is denoted as B; In step four, when comparing the length, width, and height values, if A = B, it is considered that the length, width, and height values match; In step seven, when comparing the length, width, and height values, if Z ∈ A, an element m that is non-Z and not greater than X can be found in set A, and an element n that is non-Z, non-n, and not greater than Y can be found in set A, then it is considered that the length, width, and height values match.

5. The automatic acquisition and verification method for the inbound and outbound information of regular packaged goods according to claim 4, wherein: In step four, when comparing the length, width, and height values, sort the elements in set A from largest to smallest or from smallest to largest, then sort the elements in set B in the same sorting manner, and then compare the sizes of the elements in set A and set B one by one in order. If the sizes of all elements in the two sets are the same, it is considered that the length, width, and height values match; In step seven, when comparing the length, width, and height values, compare Z with the elements in set A one by one. If the value is the same as one of the elements, denote the elements in set A other than Z as a and b; if X is not less than a and Y is not less than b, or X is not less than b and Y is not less than a, then it is considered that the length, width, and height values match.

6. The automatic acquisition and verification method for the inbound and outbound information of regular-packaged goods according to claim 2, wherein: The robotic arm (7) grabs the regular packaged goods (2) through the gripper. A point is selected between the claws of the gripper as the gripper center, and the distance from the gripper center to each claw is the same. The robotic arm (7) is equipped with a positioning device for real-time positioning of the gripper center coordinates. The coordinates of the intersection point of the depth control edge and the offset control edge are denoted as (0, 0, 0), and the coordinates of the geometric center of the regular packaged goods (2) are denoted as (X / 2, Y / 2, Z / 2); When grabbing the regular packaged goods (2), the opening width of the gripper is greater than X, and the gripper center is located between the geometric center of the regular packaged goods (2) and the positioning surface and lower than the geometric center of the regular packaged goods (2); In step four, when the gripper just grabs the regular packaged goods (2), record the gripper center coordinates (x1, y1, z1) at this time; In step seven, when the regular packaged goods (2) are placed into the detection channel (1), if the gripper center reaches (x1, y1, z1), it is regarded that the position and posture of the regular packaged goods (2) are qualified, and the gripper can be released.

7. The automatic acquisition and verification method for the inbound and outbound information of regular packaged goods according to claim 1, wherein: The automatic acquisition and verification method uses multiple detection channels (1) and multiple sets of robotic arms (7) to enable the inbound and outbound of the regular packaged goods (2) stored in the warehouse and on the shelves to be carried out simultaneously.

Images