Search

CN-122022675-A - Intelligent inspection and compliance assessment method and system for commodity display

CN122022675ACN 122022675 ACN122022675 ACN 122022675ACN-122022675-A

Abstract

The invention discloses a commodity display intelligent inspection and compliance assessment method and system, wherein the method comprises the steps of acquiring multi-mode data of a goods shelf area through a depth camera and radio frequency reading equipment which are mounted on a mobile platform, and generating an original inspection data stream; the method comprises the steps of carrying out local feature enhancement processing based on an attention mechanism on an original inspection data stream to construct a shelf display digital snapshot, carrying out multi-dimensional dynamic matching on the shelf display digital snapshot and a preset display rule base, calculating position compliance, plumpness and price tag consistency indexes by combining time dimension sales data, outputting a multi-dimensional compliance analysis result, and mining and automatically outputting a display optimization suggestion list according to association rules based on the multi-dimensional compliance analysis result. By utilizing the embodiment of the invention, the accuracy, the efficiency and the decision support capability of commodity display compliance detection can be improved.

Inventors

  • Feng Baodan
  • WANG HAIYONG

Assignees

  • 杭州吉智云信息技术有限公司

Dates

Publication Date
20260512
Application Date
20260120

Claims (10)

  1. 1. An intelligent inspection and compliance assessment method for commodity display, which is characterized by comprising the following steps: acquiring multi-mode data of a goods shelf area through a depth camera and radio frequency reading equipment which are mounted on a mobile platform, and generating an original inspection data stream containing image point cloud and commodity electronic codes; Carrying out local feature enhancement processing based on an attention mechanism on the original inspection data stream, extracting visual features of commodity packages, and setting orientation and space occupation relation, and constructing a shelf display digital snapshot; carrying out multidimensional dynamic matching on the shelf display digital snapshot and a preset display rule base, calculating position compliance, plumpness and price tag consistency indexes by combining time dimension sales data, and outputting a multidimensional compliance analysis result; Based on the multidimensional compliance analysis result, generating a visual evaluation report through three-dimensional scene reconstruction and difference labeling, and mining according to association rules to automatically output a display optimization suggestion list.
  2. 2. The method of claim 1, wherein the acquiring multi-modal data of the shelf area by the depth camera and the rf reader device onboard the mobile platform generates an original inspection data stream comprising the image point cloud and the commodity electronic code, comprising: Controlling the mobile platform to uniformly travel along a preset inspection route, synchronously triggering the depth camera to acquire a continuous depth image sequence of a goods shelf area, and simultaneously scanning electronic codes of goods through radio frequency reading equipment to generate an original depth image sequence and an electronic code data stream which are synchronously acquired; Performing interframe registration and noise filtering processing on the original depth image sequence, and fusing a plurality of frames of depth images by adopting an iterative nearest point algorithm to generate dense three-dimensional point cloud, so as to generate a three-dimensional point cloud model of a shelf area; Carrying out space-time alignment on the electronic coding data stream and the three-dimensional point cloud model, and establishing a corresponding region of each electronic code in the point cloud through timestamp matching and space mapping to generate a code-point cloud mapping relation table; And integrating the three-dimensional point cloud model and the code-point cloud mapping relation table, constructing a fusion data stream containing space geometric information and commodity identification information, and finally generating an original inspection data stream containing image point cloud and commodity electronic codes.
  3. 3. The method of claim 2, wherein the performing attention mechanism-based local feature enhancement processing on the original inspection data stream, extracting visual features of the commodity package, placing orientation, and space occupation relation, and constructing a shelf display digital snapshot comprises: Extracting a three-dimensional point cloud model from an original inspection data stream, dividing the point cloud into a plurality of local areas by adopting a voxel grid segmentation algorithm, and generating a local point cloud block set; Applying an attention-enhanced feature extraction network to each local point cloud block, wherein the network focuses on key surface areas of commodity packages through a self-attention mechanism, extracts color textures, bar code positions and brand identification features, and generates an enhanced visual feature vector; calculating the space orientation parameter of each commodity based on the enhanced visual feature vector, determining the main axis direction of commodity package by adopting a principal component analysis algorithm, estimating the included angle between the commodity and the plane of the goods shelf, and generating a commodity placement orientation matrix; And integrating the enhanced visual feature vectors of all the local blocks and the commodity placement orientation matrix, constructing a topological relation diagram describing the relative positions and the spatial layout of the commodities, and finally generating a shelf display digital snapshot.
  4. 4. The method of claim 3, wherein the step of dynamically matching the digitized snap of the shelf display with a preset display rule base in a multi-dimensional manner, calculating a position compliance index, a plumpness index and a price tag compliance index in combination with the time dimension sales data, and outputting a multi-dimensional compliance analysis result comprises: Analyzing commodity distribution information in the shelf display digital snapshot, extracting display standard templates of corresponding classes from a display rule base, wherein the display standard templates comprise commodity placement positions, layer height requirements and adjacent commodity combination rules, and generating standard display templates; Carrying out multidimensional matching on the digital snapshot and the standard display template, calculating the structural similarity between the actual commodity layout and the standard template by adopting a graph matching algorithm, and generating a layout matching degree score; The historical sales frequency and the backorder records of all business grades are analyzed by combining the time dimension sales data, the utilization efficiency of shelf space and the timely index of replenishment are calculated, and a sales data association analysis result is generated; And comprehensively calculating the position compliance, the goods shelf fullness and the price tag consistency three-dimensional compliance indexes based on the layout matching degree score and the sales data association analysis result, and finally outputting a multi-dimensional compliance analysis result.
  5. 5. The method of claim 4, wherein generating a visual assessment report through three-dimensional scene reconstruction and difference labeling based on the multi-dimensional compliance analysis result, and automatically outputting a list of showcase optimization suggestions according to association rule mining, comprises: Marking a problem area in a three-dimensional point cloud model of an original inspection data stream according to unqualified items in a multi-dimensional compliance analysis result, distinguishing different violation types by adopting color coding, and generating a three-dimensional scene model with difference labeling; generating a multi-view rendering image based on the three-dimensional scene model with the difference annotation, wherein the multi-view rendering image comprises a top view, a side view and a local close-up view, and adding the rule index data annotation to generate a visual evaluation chart; the symbiotic relation among the violation patterns is analyzed by applying an association rule mining algorithm, and violation combinations and potential reasons thereof which frequently and simultaneously occur are found, so that a violation pattern association rule set is generated; Specific adjustment suggestions and operation steps are generated for each violation pattern by combining the visual evaluation chart, the violation pattern association rule set and the standard display template, and finally a complete evaluation document containing visual evaluation reports and display optimization suggestion lists is output.
  6. 6. An intelligent inspection and compliance assessment system for merchandise display, the system comprising: The acquisition module is used for acquiring multi-mode data of a goods shelf area through a depth camera and radio frequency reading equipment which are carried on the mobile platform, and generating an original inspection data stream containing image point cloud and commodity electronic codes; the extraction module is used for carrying out local feature enhancement processing based on an attention mechanism on the original inspection data stream, extracting visual features of commodity packages, placing orientation and space occupation relation, and constructing a shelf display digital snapshot; The matching module is used for carrying out multidimensional dynamic matching on the shelf display digital snapshot and a preset display rule base, calculating position compliance, plumpness and price tag consistency indexes by combining time dimension sales data, and outputting a multidimensional compliance analysis result; The evaluation module is used for generating a visual evaluation report through three-dimensional scene reconstruction and difference labeling based on the multi-dimensional compliance analysis result, and automatically outputting a display optimization suggestion list according to association rule mining.
  7. 7. The system according to claim 6, wherein the acquisition module is specifically configured to: Controlling the mobile platform to uniformly travel along a preset inspection route, synchronously triggering the depth camera to acquire a continuous depth image sequence of a goods shelf area, and simultaneously scanning electronic codes of goods through radio frequency reading equipment to generate an original depth image sequence and an electronic code data stream which are synchronously acquired; Performing interframe registration and noise filtering processing on the original depth image sequence, and fusing a plurality of frames of depth images by adopting an iterative nearest point algorithm to generate dense three-dimensional point cloud, so as to generate a three-dimensional point cloud model of a shelf area; Carrying out space-time alignment on the electronic coding data stream and the three-dimensional point cloud model, and establishing a corresponding region of each electronic code in the point cloud through timestamp matching and space mapping to generate a code-point cloud mapping relation table; And integrating the three-dimensional point cloud model and the code-point cloud mapping relation table, constructing a fusion data stream containing space geometric information and commodity identification information, and finally generating an original inspection data stream containing image point cloud and commodity electronic codes.
  8. 8. The system according to claim 7, wherein the extraction module is specifically configured to: Extracting a three-dimensional point cloud model from an original inspection data stream, dividing the point cloud into a plurality of local areas by adopting a voxel grid segmentation algorithm, and generating a local point cloud block set; Applying an attention-enhanced feature extraction network to each local point cloud block, wherein the network focuses on key surface areas of commodity packages through a self-attention mechanism, extracts color textures, bar code positions and brand identification features, and generates an enhanced visual feature vector; calculating the space orientation parameter of each commodity based on the enhanced visual feature vector, determining the main axis direction of commodity package by adopting a principal component analysis algorithm, estimating the included angle between the commodity and the plane of the goods shelf, and generating a commodity placement orientation matrix; And integrating the enhanced visual feature vectors of all the local blocks and the commodity placement orientation matrix, constructing a topological relation diagram describing the relative positions and the spatial layout of the commodities, and finally generating a shelf display digital snapshot.
  9. 9. A storage medium having a computer program stored therein, wherein the computer program is arranged to perform the method of any of claims 1-5 when run.
  10. 10. An electronic device comprising a memory and a processor, wherein the memory has stored therein a computer program, the processor being arranged to run the computer program to perform the method of any of claims 1-5.

Description

Intelligent inspection and compliance assessment method and system for commodity display Technical Field The invention belongs to the technical field of commodity data processing, and particularly relates to an intelligent inspection and compliance assessment method and system for commodity display. Background Currently, the retail industry relies mainly on manual visual or fixed camera-based image recognition methods for compliance inspection of merchandise displays. The traditional means have the problems of low efficiency, strong subjectivity, single coverage dimension and the like, and the multi-mode accurate sensing and dynamic evaluation of the shelf state are difficult to realize. Although some schemes attempt to introduce RFID or image recognition technology, the comprehensive analysis capability of commodity placement orientation, spatial relationship and linkage with sales data is generally lacking, so that the inspection result is one-sided and feedback is lagged, and real-time and intelligent display optimization decision cannot be effectively supported. In addition, the existing method is limited to two-dimensional visual comparison, has obvious defects in three-dimensional space compliance judgment, implicit rule mining and visual interaction, and restricts the improvement of the refinement level of retail operation. Disclosure of Invention The invention aims to provide an intelligent inspection and compliance assessment method and system for commodity display, which are used for solving the defects in the prior art and improving the accuracy, efficiency and decision support capability of commodity display compliance detection. One embodiment of the application provides a commodity display intelligent patrol and compliance assessment method, which comprises the following steps: acquiring multi-mode data of a goods shelf area through a depth camera and radio frequency reading equipment which are mounted on a mobile platform, and generating an original inspection data stream containing image point cloud and commodity electronic codes; Carrying out local feature enhancement processing based on an attention mechanism on the original inspection data stream, extracting visual features of commodity packages, and setting orientation and space occupation relation, and constructing a shelf display digital snapshot; carrying out multidimensional dynamic matching on the shelf display digital snapshot and a preset display rule base, calculating position compliance, plumpness and price tag consistency indexes by combining time dimension sales data, and outputting a multidimensional compliance analysis result; Based on the multidimensional compliance analysis result, generating a visual evaluation report through three-dimensional scene reconstruction and difference labeling, and mining according to association rules to automatically output a display optimization suggestion list. Yet another embodiment of the present application provides a merchandise display intelligent patrol and compliance assessment system, the system comprising: The acquisition module is used for acquiring multi-mode data of a goods shelf area through a depth camera and radio frequency reading equipment which are carried on the mobile platform, and generating an original inspection data stream containing image point cloud and commodity electronic codes; the extraction module is used for carrying out local feature enhancement processing based on an attention mechanism on the original inspection data stream, extracting visual features of commodity packages, placing orientation and space occupation relation, and constructing a shelf display digital snapshot; The matching module is used for carrying out multidimensional dynamic matching on the shelf display digital snapshot and a preset display rule base, calculating position compliance, plumpness and price tag consistency indexes by combining time dimension sales data, and outputting a multidimensional compliance analysis result; The evaluation module is used for generating a visual evaluation report through three-dimensional scene reconstruction and difference labeling based on the multi-dimensional compliance analysis result, and automatically outputting a display optimization suggestion list according to association rule mining. A further embodiment of the application provides a storage medium having a computer program stored therein, wherein the computer program is arranged to perform the method of any of the preceding claims when run. Yet another embodiment of the application provides an electronic device comprising a memory having a computer program stored therein and a processor configured to run the computer program to perform the method recited in any of the preceding claims. Compared with the prior art, the intelligent inspection and compliance assessment method for commodity display can improve accuracy, efficiency and decision support capability of commodity display compliance d