A method and device for automatic detection of shelf safety based on YOLOv7

Through the automatic shelf safety detection method based on YOLOv7, the neighborhood variance modeling and improved YOLOv7 network structure are used, combined with apriltag positioning technology, human errors and high cost problems in existing shelf detection are solved, and the accurate and real-time identification and positioning of shelf defects are achieved.

CN117237835BActive Publication Date: 2025-07-22YANAN UNIV
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202310115581.2
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2023-02-14
Publication Date
2025-07-22
Estimated Expiration
2043-02-14

AI Technical Summary

Technical Problem

The existing shelf inspection methods rely on manual inspection, which has problems of human error and high cost, and the sensor solution requires multiple sensors and the sensitivity is affected by distance.

Method used

The automatic shelf safety detection method based on YOLOv7 is adopted, and the data set is processed through the neighborhood variance modeling mechanism, and the lighting model is constructed for data expansion. The improved YOLOv7 network structure is used to deploy it on NVIDIA Jetson Nano, and automated detection and positioning are achieved in combination with apriltag positioning technology.

Benefits of technology

It realizes accurate and real-time identification and positioning of shelf defects, improves detection accuracy and speed, and reduces human errors and costs.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN117237835B_ABST
    Figure CN117237835B_ABST
Patent Text Reader

Abstract

The present invention provides a method and device for automatic detection of shelf safety based on YOLOv7, which relates to the field of detection technology. The method includes: obtaining an original data set. After processing the original data set through a neighborhood variance modeling mechanism, the data obtained through the neighborhood variance modeling mechanism is labeled to form labels, and a first data set is made. The first data set is divided into a training set, a validation set, and a test set according to a preset ratio. The training set is imported into an improved YOLOv7 network structure, and a rack structure defect recognition model is trained. The rack structure defect recognition model is converted using a TensorRT engine and deployed on an NVIDIA Jetson Nano, and the defective rack is located through an AprilTag positioning technology, completing the accurate and real-time recognition and positioning of the pallet rack mechanism defects, thereby realizing the automatic detection and positioning of shelf safety defects.
Need to check novelty before this filing date? Find Prior Art

Citation Information

Patent Citations

  • Insulator defect detection method based on improved YOLOv3 convolutional neural network

    CN114283117A

  • Small transparent plastic product defect detection method based on YOLOv5

    CN115147380A