Method and system for object position identification and rapid positioning and grabbing of retail shelf

Abstract

The invention relates to the field of intelligent retail, and discloses a retail shelf object position identification and rapid positioning grabbing method and system, and the method comprises the following steps: S1, obtaining the initial position data of an infrared distance measurement sensor and an angle control device which are disposed in front of a shelf, and building a three-dimensional coordinate system employing the infrared distance measurement sensor as a reference point; s2, controlling an infrared distance measuring sensor to be aligned with an initial position point of a target layer of the goods shelf, and calculating a space coordinate of a reference point of the layer after obtaining corresponding angle parameters and distance measuring data; s3, dividing the space into a plurality of discrete position points according to the horizontal length of the goods shelf and preset positioning precision; and S4, constructing a data structure for representing the state of the discrete position point object. According to the method, the space in front of the goods shelf is divided into the equidistant scattered position points, and the bit array is constructed, so that minimum data compression and structured expression of object distribution information are realized, and the effects of extremely low memory occupation, high operation efficiency and adaptation to an embedded system are achieved.

Description

Technical Field

[0001] The present invention relates to the field of intelligent retail, and specifically to a method and system for object position recognition and rapid positioning and grasping of retail shelves. Background Art

[0002] With the continuous development of the intelligent retail industry, its market scale has been increasing year by year. As the core infrastructure in the retail system, retail shelves are also constantly undergoing technological innovation and function upgrade. Driven by the improvement of consumer experience, the optimization of operation efficiency and the trend of industry intelligence, retail shelves are gradually evolving from traditional static display to intelligent perception and dynamic response. Especially in scenarios such as large supermarkets, convenience stores and unmanned retail terminals, where there are a wide variety of products and dense placement, how to achieve rapid positioning and accurate recognition of products on the shelves has become a key factor affecting the overall operation efficiency and customer experience.

[0003] Currently, the object recognition and position location of retail shelves mainly rely on perception and positioning technologies such as image recognition, lidar scanning, millimeter wave radar or UWB (Ultra Wide Band). Although these technologies have certain recognition capabilities to a certain extent, there are many limitations. For example, image recognition is easily affected by changes in lighting, occlusion and object similarity, and the recognition accuracy is unstable; although lidar and millimeter wave radar can provide relatively accurate spatial information, their hardware costs are high, and they are large in size and high in power consumption, which are not suitable for flexible deployment in small and medium-sized retail terminals; although UWB has high accuracy, the system integration is complex, relying on multiple tags and base stations, and is sensitive to environmental interference at the same time. In addition, most of the above technologies rely on the processing and fusion of large-scale data, requiring high-performance processors and complex algorithms, which are difficult to be efficiently implemented in edge computing environments or resource-constrained devices. Summary of the Invention

[0004] Aiming at the deficiencies of the prior art, the present invention provides a method and system for object position recognition and rapid positioning and grasping of retail shelves, which solves the problems of high-cost sensors or complex image processing for object positioning in existing retail shelves, difficult system deployment and large computational overhead.

[0005] To achieve the above objectives, the present invention is realized through the following technical solutions: A method for object position recognition and rapid positioning and grasping of retail shelves, including the following steps:

[0006] S1: Obtain the initial position data of the infrared ranging sensor and the angle control device installed in front of the shelf, and establish a three-dimensional coordinate system with the infrared ranging sensor as the reference point;

[0007] S2: Control the infrared ranging sensor to align with the starting position point of the target layer of the shelf. After obtaining the corresponding angle parameters and ranging data, calculate the spatial coordinates of the reference benchmark point of this layer;

[0008] S3: Divide the space into multiple discrete position points according to the horizontal length of the shelf and the preset positioning accuracy;

[0009] S4: Construct a data structure for representing the object states at the discrete position points;

[0010] S5: Compare the actual ranging values at the discrete position points with the theoretical thresholds to determine whether there are objects at each position point, and record the corresponding marking information in the data structure;

[0011] S6: Repeat the above steps to scan and identify multiple target levels of the shelf to form a two-dimensional position matrix;

[0012] S7: Receive the identification information of the target object, obtain the coordinate data corresponding to the target position point according to the two-dimensional position matrix, and generate a control instruction for the grasping device to perform the object grasping operation.

[0013] Preferably, the infrared ranging sensor is installed on the top of the angle control device, and the angle control device includes two independent rotation axes for respectively controlling the angle changes of the infrared ranging sensor in the horizontal and vertical directions.

[0014] Preferably, in the three-dimensional coordinate system, the x-axis is the forward direction directly facing the shelf by the infrared ranging sensor, the z-axis is the vertical direction directly facing the ground, and the y-axis is the horizontal axis perpendicular to the x-axis and the z-axis.

[0015] Preferably, the data structure is a bit array structure, and its length is jointly determined by the horizontal length of the shelf and the preset positioning accuracy. If the total number of bits is less than one whole byte, the last bit is padded with empty spaces.

[0016] Preferably, S3 includes:

[0017] Divide the horizontal length of the shelf by the positioning accuracy to obtain the number of division steps;

[0018] Encode the number of division steps in binary form and convert it into a byte structure for storage.

[0019] Preferably, when the byte structure result in S3 is less than one whole byte, 0 is padded at the end to fill it to a complete byte for data alignment.

[0020] Preferably, in S6, the process of combining the data structures of each level to form a two-dimensional position matrix includes: stacking and arranging the bit arrays corresponding to each layer in the order of their vertical levels in the shelf to construct a two-dimensional array containing information of multiple layers of shelves.

[0021] Preferably, S7 includes:

[0022] After receiving the target object identification information, the spatial coordinate data corresponding to the object is obtained by querying the matching bit marker positions in the two-dimensional position matrix, and a control command is generated with the coordinates as the target position of the grasping device.

[0023] Preferably, in S2, the spatial coordinates of the reference reference point of the target layer are jointly calculated by the distance value obtained by the infrared ranging sensor and the angle parameters of the angle control device, and the angle parameters include the pitch angle and the horizontal rotation angle.

[0024] An object position recognition and rapid positioning and grasping system for a retail shelf, comprising:

[0025] An infrared ranging sensor for collecting distance data at different angles in front of the shelf;

[0026] An angle control device for controlling the infrared ranging sensor to rotate in the horizontal and vertical directions;

[0027] A controller for receiving the ranging data and angle information and calculating the spatial coordinates of multiple reference points;

[0028] A storage module for constructing and storing a bit array and a two-dimensional position matrix;

[0029] An actuator for controlling the grasping device to complete object grasping according to the target position.

[0030] The present invention provides a method and a system for object position recognition and rapid positioning and grasping of a retail shelf.

[0031] Has the following beneficial effects:

[0032] 1. By dividing the space in front of the shelf into equally spaced discrete position points and constructing a bit array to store the object states of each position point at the bit level, the present invention realizes extremely small data compression and structured expression of the object distribution information, obtaining the effects of extremely low memory occupancy, high operation efficiency, and adaptation to embedded systems.

[0033] 2. By setting a single infrared ranging sensor and combining it with a biaxial angle control device for multi-angle scanning, the present invention realizes coverage ranging of multiple positions of the shelf, without the need for multi-sensor collaboration and complex calibration, obtaining the effects of simple structure, fast layout, and significantly reduced deployment cost.

[0034] 3. By collecting angle control parameters and ranging values to construct spatial coordinate points and combining the theoretical path difference to judge the object occlusion state, the present invention realizes highly reliable judgment of the presence or absence of a target without image recognition, obtaining the effects of strong anti-light interference, good real-time performance, and more stable target determination. Description of the Drawings

[0035] Figure 1 Schematic diagram of the step structure of the present invention;

[0036] Figure 2 System module architecture diagram of the present invention. Specific implementation mode

[0037] The technical solutions of the present invention will be clearly and completely described below in conjunction with the accompanying drawings of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0038] Embodiment 1:

[0039] Please refer to the attached Figure 1 , the embodiment of the present invention provides a method for object position recognition and rapid positioning and grasping of a retail shelf, including the following steps:

[0040] S1: Obtain the initial position data of the infrared ranging sensor and the angle control device installed in front of the shelf, and establish a three-dimensional coordinate system with the infrared ranging sensor as the reference point; the infrared ranging sensor is installed on the top of the angle control device, and the angle control device includes two independent rotation axes, which are respectively used to control the angle change of the infrared ranging sensor in the horizontal direction and the vertical direction.

[0041] Specifically, this step first needs to obtain the initial position data of the infrared ranging sensor and the angle control device installed in front of the shelf. For this purpose, the infrared ranging sensor is fixed on the top of the angle control device, and the angle control device includes two independent rotation axes, which are respectively used to control the angle change of the infrared ranging sensor in the horizontal direction and the vertical direction. The combination of the sensor and the angle control device provides flexible scanning capabilities, and can adjust the angle of the sensor according to needs, so as to perform high-precision measurements on different areas of the shelf.

[0042] Once the device is installed, the system will record the position data of the infrared ranging sensor and establish a three-dimensional coordinate system with the sensor position as the reference point. In this coordinate system, the installation position of the infrared ranging sensor is used as the origin, and the coordinate axes are defined as: the horizontal direction is the X axis, the vertical direction is the Z axis, and the depth direction is the Y axis. Through such a coordinate system setting, the spatial relationship between the sensor and the shelf can be accurately calibrated to ensure the accuracy and consistency of subsequent measurement data.

[0043] The function of the angle control device is to ensure that the sensor can scan within different angular ranges by adjusting the angle of the sensor. Specifically, the rotation axis in the horizontal direction controls the sensor to rotate horizontally, and the rotation axis in the vertical direction controls the scanning angle of the sensor in the vertical direction. In this way, the infrared ranging sensor can cover all levels of the shelf, capture the position data of the objects on each level, and ensure that no items are missed.

[0044] During the measurement process, the sensor will continuously adjust its angle, gradually scan the objects at different positions on the shelf, and record the distance data of each measurement point in real time. These measurement data will be used to calculate the spatial coordinates of the objects, providing data support for subsequent positioning and rapid grasping. In this way, the system can adapt to shelves of different sizes and shapes without large-scale hardware changes, and can flexibly adjust the scanning angle to meet various specific requirements.

[0045] S2: Control the infrared ranging sensor to align with the starting position point of the target layer of the shelf. After obtaining the corresponding angle parameters and ranging data, calculate the spatial coordinates of the reference benchmark point of this layer; in the three-dimensional coordinate system, the x-axis is the positive front direction facing the shelf by the infrared ranging sensor, the z-axis is the vertical direction facing the ground, and the y-axis is the horizontal axis perpendicular to the x-axis and the z-axis. In step S2, the spatial coordinates of the reference benchmark point of the target layer are jointly calculated by the distance value obtained by the infrared ranging sensor and the angle parameters of the angle control device. The angle parameters include the pitch angle and the horizontal rotation angle.

[0046] Specifically, first, the system controls the infrared ranging sensor to align with the starting position point of the target layer of the shelf. This position point is set based on the shelf structure and the adjustment accuracy of the angle control device. Specifically, the infrared ranging sensor is adjusted to a specific angle through the angle control device to ensure that the sensor can accurately align with the position of the target layer on the shelf. At this time, the angle control device will obtain the corresponding angle parameters, including the pitch angle and the horizontal rotation angle, which are the key parameters to determine the spatial relationship between the sensor and the target layer.

[0047] After aligning with the starting position, the infrared ranging sensor will start measuring the specific position of the target layer. By sending infrared signals and receiving the reflected signals, the ranging sensor can calculate the distance from the sensor to the objects on the target layer. This distance data, together with the angle parameters, is used for subsequent spatial coordinate calculations. To ensure the measurement accuracy, the cooperation between the infrared ranging sensor and the angle control device is very important. The ranging results and angle adjustments of the sensor can provide accurate positioning information on shelves of different levels.

[0048] Next, by combining the measured distance data with the angle parameters, the system calculates the spatial coordinates of the reference benchmark point on the target layer. The three-dimensional coordinate system is set as follows: the x-axis points to the directly forward direction of the infrared distance sensor facing the shelf, the z-axis points to the vertically upward direction, parallel to the ground, and the y-axis is perpendicular to the x-axis and the z-axis, forming a right-handed coordinate system. In this coordinate system, the spatial coordinates of the reference benchmark point on the target layer can be calculated through a formula. With the known distance value and angle parameters, the position of the reference benchmark point on the target layer in the three-dimensional space can be accurately deduced.

[0049] The specific calculation process is as follows: According to the pitch angle (θ_vertical) and the horizontal rotation angle (θ_horizontal), as well as the distance measurement value (S) obtained by the infrared distance sensor, combined with the setting of the three-dimensional coordinate system, trigonometric functions are used for calculation. The pitch angle controls the position change of the sensor in the vertical direction, and the horizontal rotation angle affects the angle change of the sensor in the horizontal direction. By combining these angle parameters with the distance value, the accurate coordinates of the reference benchmark point can be obtained.

[0050] S3: Divide the space into multiple discrete position points according to the horizontal length of the shelf and the preset positioning accuracy; including: dividing the horizontal length of the shelf by the positioning accuracy to obtain the number of division steps; encoding the number of division steps in binary form and converting it into a byte structure for storage. When the byte structure result is less than a full byte, 0s are filled at the end to complete the byte for data alignment.

[0051] Specifically, first divide the horizontal length of the shelf by the preset positioning accuracy to calculate the number of division steps required. This number of steps represents the number of discrete position points after division within the horizontal length of the shelf according to the set accuracy. For example, if the horizontal length of the shelf is 1.5 meters and the positioning accuracy is 5 centimeters, then the number of division steps will be 1.5 meters divided by 5 centimeters, that is, 30 discrete position points.

[0052] Then, the system encodes these division steps in binary form. Binary encoding can concisely represent these discrete position points, ensuring efficient storage and processing of data. The encoded binary data is converted into binary digits according to the number of division steps, representing the state (with or without items) of each position point. These binary digits will form a long binary sequence.

[0053] Next, the encoded binary data will be converted into a byte structure for storage. Since computer memory usually stores data in bytes as the basic unit, to ensure that the data can be byte-aligned, the system needs to divide the binary sequence into bytes. Each byte contains 8 bits. Therefore, if the result of binary encoding is less than a full byte (i.e., the number of bits is less than 8), the system will fill 0s at the end until the requirement of a full byte is met. The purpose of this is to ensure memory alignment during data storage and avoid memory waste or performance issues caused by improper storage.

[0054] The process of supplementing 0 for data alignment is usually a common operation during the storage process of a computer. It can improve the efficiency of data access. By supplementing 0s to fill to the byte boundary, the storage system can perform data access operations more efficiently, while avoiding the generation of fragmented memory and improving the utilization rate of memory. In this way, not only the integrity of the data is guaranteed, but also the entire system is more efficient in subsequent data processing.

[0055] S4: Construct a data structure for representing the state of an object at discrete position points; the data structure is a bit array structure, and its length is jointly determined by the horizontal length of the shelf and the preset positioning accuracy. If the total number of bits is less than a full byte, the last bit fills the vacant position.

[0056] Specifically, first determine the length of the bit array. According to the horizontal length of the shelf and the preset positioning accuracy, the system will calculate the total number of discrete position points that need to be divided on the shelf. This total number determines the length of the bit array. Each position point is represented by a bit (0 or 1), indicating whether there is an object at that position. Specifically, the horizontal length of the shelf divided by the positioning accuracy gives the number of division steps, and each position point corresponds to a bit. During the calculation process, if the number of discrete position points divided is large, a bit array of the corresponding length needs to be constructed based on these points.

[0057] Then, the system constructs a bit array data structure according to the number of division steps. Each position point is assigned a binary value, 0 indicating that there is no object at that position, and 1 indicating that there is an object at that position. Each bit in the bit array corresponds to a specific position on the shelf, ensuring that the state of each object (whether there is an object) is accurately recorded.

[0058] Next, considering that computer storage usually stores data in units of bytes (8 bits), after the length of the bit array is determined, if the total number of bits is less than a complete byte (i.e., the total number of bits is not a multiple of 8), then the bit array needs to be padded. The system will append empty bits 0 at the end of the bit array until the length of the array meets the requirement of a whole byte. This is to ensure that data can be byte-aligned during storage, thereby improving storage efficiency and access speed. The operation of padding 0 ensures that there will be no problems with the array during storage due to the lack of a complete byte and also avoids the generation of memory fragmentation.

[0059] The design of this data structure has high storage efficiency. By using a bit array, the system can represent the state of objects on the shelf in a very compact form while effectively reducing memory usage. Each bit represents the position state of an object, which is both concise and intuitive, facilitating subsequent calculations and data processing.

[0060] S5: Compare the actual ranging values of discrete position points with the theoretical threshold to determine whether there are objects at each position point and record the corresponding marking information in the data structure;

[0061] Specifically, in the previous steps, the actual ranging values of each discrete position point have been obtained through an infrared ranging sensor. These ranging values represent the actual distances from the infrared ranging sensor to the objects on the shelf. In this step, the system will compare these actual ranging values with the theoretical threshold, and the theoretical threshold represents the maximum distance at which an object can be measured or the expected placement range of objects on a certain layer.

[0062] Specifically, the system will first set a theoretical threshold, which can be set according to the structure of the shelf, the size of the items, and the measurement accuracy requirements. For example, if the system needs to detect whether the items on the shelf are in the specified positions, the theoretical threshold may be set to the maximum size of the items or the measurement accuracy of the ranging sensor. During the comparison process, if the actual ranging value is less than or equal to the theoretical threshold, it is considered that there is an object at that position point; if the actual ranging value is greater than the theoretical threshold, it is considered that there is no object at that position point.

[0063] Next, based on the comparison results, the system records the marking information of the corresponding position points in the data structure. The previously mentioned data structure is a bit array, and the system will represent the state of each discrete position point with a single bit (0 or 1). If a certain position point is determined to have an object, the system will record 1 in the corresponding position of the bit array, indicating that there is an object at that position; if that position point is determined to have no object, the system will record 0, indicating that the position is empty.

[0064] The key to this process is to judge the objects on the shelf by comparing the ranging values with the theoretical thresholds. In this way, the system can accurately identify the distribution of objects on the shelf according to the actual ranging values of the sensors and record the results in the form of a bit array. This method is efficient and accurate, and can complete the large-scale object recognition in a short time to meet the requirements of rapid item search and positioning in the retail environment.

[0065] In addition, through this method, the system can effectively avoid misjudgments or omissions in traditional image recognition or other high-cost and high-complexity technologies, ensuring the accuracy of positioning and recognition. The comparison operation is simple and efficient, and can meet the requirements of object recognition on multi-level and various types of shelves, providing support for subsequent object grasping or repositioning.

[0066] S6: Repeat the above steps to scan and identify multiple target levels of the shelf to form a two-dimensional position matrix; in S6, the process of combining the data structures of each level to form a two-dimensional position matrix includes: stacking the bit arrays corresponding to each level in the vertical level order in the shelf to construct a two-dimensional array containing multi-level shelf information.

[0067] Specifically, according to the previous steps, each layer of the shelf is scanned and identified to obtain the bit array of each layer. Each bit array corresponds to discrete position points on a certain layer of the shelf, and each bit of each bit array represents whether there is an object at that position point. If there is an object at that position, the corresponding bit in the array is 1; if there is no object at that position, the bit is 0. In this way, the system can record the specific distribution of objects on each layer of the shelf.

[0068] Next, stack these bit arrays in the vertical level order of the shelf. The bit array of each layer of the shelf represents the position state of the objects on that layer. The stacking operation is to combine the bit arrays of each layer together in order to form a unified two-dimensional data structure. When stacking, the bit array of the first layer is located in the first row of the matrix, the bit array of the second layer is located in the second row, and so on. In this way, each row corresponds to a layer of the shelf, and the columns correspond to discrete position points at different positions.

[0069] Finally, by stacking the bit arrays of each layer of the shelf, a two-dimensional array containing multi-level shelf information is constructed. Each row of the two-dimensional matrix represents a level of the shelf, and each column in the matrix corresponds to a position point on the shelf. This data structure can comprehensively reflect the object distribution of the entire shelf, facilitating subsequent processing and analysis.

[0070] For example, assume that the shelf has three layers and each layer has five position points. Then the final two-dimensional matrix will be a 3-row and 5-column array, where each row contains the status of all position points on that layer (1 indicates the presence of an object, and 0 indicates the absence of an object). This data structure can accurately reflect the distribution of objects on each layer of the shelf and is convenient for further positioning and item identification.

[0071] S7: Receive the identification information of the target object, obtain the coordinate data corresponding to the target position point according to the two-dimensional position matrix, and generate a control instruction for the grasping device to perform the object grasping operation.

[0072] Specifically, after generating the complete position matrix, the system will extract the position of the target object and generate an instruction for the grasping task based on the matrix information. Each row in the matrix corresponds to a layer of the shelf, and each column represents a discrete position point in the horizontal direction. When the user or the system specifies the target item type, the system will combine the matrix status, priority strategy, or other recognition mechanisms to find the first valid position in the matrix, that is, the position with a value of 1. After finding it, the system takes the row and column indexes of this position as the preliminary identification of the target position, and then converts it into the three-dimensional grasping coordinates in the actual space through the coordinate mapping function. This mapping process is not complicated. Multiply the horizontal index by the unit step P, multiply the vertical index by the layer height ΔH, and add the origin offset to inversely calculate the space coordinates (x, y, z). If the sensor stores the coordinates of each corresponding point during the scanning process, the point coordinates corresponding to this index can be directly found from the existing point set, eliminating the need for recalculation.

[0073] After determining the target coordinates, the system transmits them to the end effector, usually a robotic arm or an electric slide rail device. The end effector plans its posture according to the three-dimensional position of this point and completes the actions of approaching, grasping, and returning at the end. The whole process does not rely on complex image recognition, nor does it require three-dimensional model reconstruction. It can complete the task only by the bit-level status matrix and simple coordinate reverse lookup.

[0074] The advantages of this approach are stability, speed, low computational overhead, and it is more suitable for edge devices or low-power scenarios. In addition, if there are multiple candidate points in the matrix, the system can also select the optimal one from multiple points according to distance, position priority, or user-configured rules to improve the flexibility of the system.

[0075] Based on the same inventive concept, the following specific embodiments are given:

[0076] Before algorithm processing, set the parameters: Assuming that the number of shelves N is 3 (N=3), the interval height of each shelf is equal, install the infrared distance sensor and angle control device near the geometric center of the front of the shelf (the infrared distance sensor is installed on the angle control device). At this time, the position of the infrared distance sensor is used as the reference point of the entire shelf. The front of the infrared sensor is the positive direction of the x-axis, the horizontal left direction is the positive direction of the y-axis, and the vertical upward direction is the positive direction of the z-axis. Taking the first shelf as an example, the specific implementation is as follows:

[0077] First, adjust the angle control device so that the infrared ranging sensor is vertically aligned with the ground and horizontally aligned with the shelf, and record the distance value (S_vertical, S_level) of the infrared ranging sensor at this time. At this time, the x, y, z coordinates of the reference point of the entire shelf are (Slevel, 0, Svertical). Then, align the infrared ranging sensor with the left column of the first shelf (all objects on the first shelf can be scanned horizontally).

[0078] Record the horizontal and vertical deflection angles (θ_level, θ_vertical) of the angle control device and the distance value (S1) of the infrared ranging sensor. Then, the coordinate values ​​x, y, and z of the reference point D1 of the object on the first shelf can be calculated as follows:

[0079] (Slevel+S1*cosθosθveral*cosθosθle, S1*cosθosθveral*sinθinθle, Svertical+S1*sinθinθveral).

[0080] After determining the reference point of the shelf, assuming that the horizontal length of the shelf is L (for example, L = 1.5 meters), the positioning accuracy is Φ (for example, Φ = 5 centimeters), and the memory size required to be stored is calculated to be M. The calculation formula for M is M = (L / Φ) / 8. If M can be rounded, the memory size required to be stored is M bytes, otherwise, the memory size required to be stored is (M+1) bytes. According to the assumed conditions, M = 4 at this time, so only 4 bytes are needed to store the position of objects on this layer.

[0081] After determining the required memory size, it is necessary to calculate the unit angle value θ of the angle control device to control the movement of the infrared sensor m1 ,θ m1 =arcsin(Φ·M·8 / S1), the angle control device is based on θ m1 The angle increment controls the detection distance of the infrared ranging sensor. Assume that the detection distance at this time is S 1c1 , the front and back placement error is T (T = 4 cm), set the storage array at this time to A1 (each A nThere are 8 bits. Since the x-axis and z-axis data of the objects on the same layer are the same, the only difference is the deviation of the y-axis. At this time, the theoretical distance value for calculating the position of this point is S 1cx , S 1cx = S1·cosθ m1 , so if S 1c1 is greater than S 1cx + T, it means there is no object placed at this position, then this bit is 0, otherwise this bit is 1, such as A1[1,0,0,0,0,0,0,0].

[0082] Calculate the corresponding S for each unit angle value θ m1 in turn according to the above steps and methods, determine the placement of all objects on the subsequent same layers until the rightmost edge of the shelf is detected. Assume that the increment when reaching the edge at this time is x, then calculate the detected distance value S 1cx at this time, S 1x , S 1x = Φ·x, and judge whether S 1x is greater than L. If S 1x is greater than L, then no further detection will be carried out, and the subsequent bits will be set to 0. Assume that the detected array list at this time is Array1, Array1 = [A1, A2,..., A M .

[0083] Repeat the above steps to detect the recognition positions of objects on subsequent different layers. For example, the coordinate values xyz of the reference reference point D2 on the second layer and the array list are: S level2 + S2*cosθ vertical2 *cosθ level2 , S2*cosθ vertical2 *sinθ level2 , S vertical2 + S2*sinθ vertical2 ), Array2 = [A1, A2,..., A M .

[0084] The coordinate values xyz of the reference reference point D3 on the third layer and the array list are: S level3 + S3*cosθ vertical3 *cosθ level3 , S3*cosθ vertical3 *sinθ level3 , S vertical3 + S3*sinθ vertical3 ), Array3 = [A1, A2,..., A M . Combine all the above data into a two-dimensional data matrix Matrix, and the content of the matrix is:

[0085] When all the position matrix data of all layers of the entire shelf are obtained through the above method, the specific position of the specific number can be quickly calculated according to the specific number, and fast positioning and grasping can be achieved.

[0086] Embodiment 2:

[0087] Please refer to the attached Figure 2 , an object position recognition and fast positioning and grasping system for a retail shelf, comprising:

[0088] An infrared ranging sensor for collecting distance data at different angles in front of the shelf;

[0089] An angle control device for controlling the rotation of the infrared ranging sensor in the horizontal and vertical directions;

[0090] A controller for receiving the ranging data and angle information and calculating the spatial coordinates of multiple reference points;

[0091] A storage module for constructing and storing a bit array and a two-dimensional position matrix;

[0092] An actuator for controlling the grasping device to complete object grasping according to the target position.

[0093] Specifically, the infrared ranging sensor is installed at the end of the angle control device and is connected to the controller through a cable for real-time collection of the ranging data in its facing direction. The angle control device consists of two independent motor units and can drive the infrared ranging sensor to rotate in the horizontal and vertical directions respectively under the command of the controller to achieve full-coverage scanning. Every time an angular displacement is completed, the infrared ranging sensor emits a ranging signal and transmits back the current distance value and sampling time, and the controller simultaneously records the current angle information and calculates the spatial coordinate point corresponding to this direction in combination with the ranging value.

[0094] These coordinate points are then uniformly sent to the storage module, which archives and manages them, and further constructs a bit array corresponding to discrete position points in space and a multi-layer two-dimensional position matrix. The bit array is used to record whether there are obstacles or target objects at each position point, and the two-dimensional matrix integrates all data by layer to form a complete object distribution map.

[0095] The controller judges the target position according to the matrix information and the task instruction, and then outputs the spatial coordinates of the target to the actuator to drive the end grasping device to perform precise grasping. The actuator is usually an electric gripper or a flexible robotic arm, and its movement path and actions are obtained by inverse kinematics of the target point position. The entire system maintains synchronous cooperation during operation. The controller is the core scheduling unit, the ranging and angle modules provide environmental perception capabilities, the storage module is responsible for status recording and historical call, and the actuator completes the final operation.

[0096] Although embodiments of the present invention have been shown and described, those of ordinary skill in the art will appreciate that various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principles and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for identifying the position of an object on a retail shelf and quickly positioning and grasping it, characterized in that, Including the following steps: S1: Obtain the initial position data of the infrared ranging sensor and the angle control device installed in front of the shelf, and establish a three-dimensional coordinate system with the infrared ranging sensor as the reference point; S2: Control the infrared ranging sensor to align with the starting position point of the target layer of the shelf. After obtaining the corresponding angle parameters and ranging data, calculate the spatial coordinates of the reference benchmark point of this layer; S3: Divide the space into multiple discrete position points according to the horizontal length of the shelf and the preset positioning accuracy; S4: Construct a data structure for representing the object state of the discrete position points; S5: Based on the comparison between the actual ranging value and the theoretical threshold of the discrete position points, determine whether there is an object at each position point, and record the corresponding marking information in the data structure; S6: Repeat the above steps to scan and identify multiple target levels of the shelf to form a two-dimensional position matrix; S7: Receive the identification information of the target object, obtain the coordinate data corresponding to the target position point according to the two-dimensional position matrix, and generate a control instruction for the gripping device to perform the object gripping operation.

2. The method for object position recognition, rapid positioning and grasping of a retail shelf according to claim 1, wherein The infrared ranging sensor is installed on the top of the angle control device, and the angle control device includes two independent rotating shafts, which are respectively used to control the angle change of the infrared ranging sensor in the horizontal direction and the vertical direction.

3. The method for object position recognition, rapid positioning and grasping of a retail shelf according to claim 1, characterized in that, In the three-dimensional coordinate system, the x-axis is the positive front direction of the infrared ranging sensor facing the shelf, the z-axis is the vertical direction facing the ground, and the y-axis is the horizontal axis perpendicular to the x-axis and the z-axis.

4. The method for object position recognition, rapid positioning and grasping of a retail shelf according to claim 1, characterized in that, The data structure is a bit array structure, and its length is jointly determined by the horizontal length of the shelf and the preset positioning accuracy. If the total number of bits is less than one whole byte, the last bit is padded with empty spaces.

5. The method for object position recognition, rapid positioning and grasping of a retail shelf according to claim 1, wherein The S3 includes: Divide the horizontal length of the shelf by the positioning accuracy to obtain the number of division steps; Encode the number of division steps in binary form and convert it into a byte structure for storage.

6. The method for object position recognition and fast positioning and grasping of a retail shelf according to claim 5, characterized in that, When the byte structure result in S3 is less than one whole byte, it is padded with 0 at the end to fill it to a complete byte for data alignment.

7. The method for identifying the position of an object on a retail shelf and quickly positioning and grasping according to claim 1, characterized in that, In S6, the process of combining the data structures of each level to form a two-dimensional position matrix includes: stacking and arranging the bit arrays corresponding to each layer in the order of their vertical levels in the shelf to construct a two-dimensional array containing information of multiple layers of shelves.

8. The method for object position recognition, rapid positioning and grasping of a retail shelf according to claim 1, wherein, The S7 includes: After receiving the target object identification information, obtain the spatial coordinate data corresponding to the object by querying the matching bit mark position in the two-dimensional position matrix, and generate a control instruction with this coordinate as the target position of the gripping device.

9. The method for object position recognition, rapid positioning and grasping of a retail shelf according to claim 1, characterized in that, In S2, the spatial coordinates of the reference benchmark point of the target layer are jointly calculated by the distance value obtained by the infrared ranging sensor and the angle parameters of the angle control device, and the angle parameters include the pitch angle and the horizontal rotation angle.

10. An object position recognition and rapid positioning and grasping system for a retail shelf, according to the method for object position recognition and rapid positioning and grasping of a retail shelf described in any one of claims 1-9, characterized in that, Including: An infrared ranging sensor for collecting distance data at different angles in front of the shelf; An angle control device for controlling the rotation of the infrared ranging sensor in the horizontal direction and the vertical direction; A controller for receiving the ranging data and angle information and calculating the spatial coordinates of multiple reference points; a storage module for constructing and storing the bit array and the two-dimensional position matrix; an actuator for controlling the gripping device to complete the object gripping according to the target position.

Images