Video monitoring warehouse in-out warehouse counting system and method based on deep learning

Abstract

The invention discloses a video monitoring warehouse in-out warehouse counting system and method based on deep learning, and relates to the technical field of deep learning. The system comprises a monitoring video acquisition module, a video processing module, an article detection and identification module, a warehouse-in and warehouse-out logic judgment module and a visual management recording module. The method comprises the following steps: S1, acquiring a monitoring video; S2, segmenting the acquired video; S3, performing manual marking; S4, respectively carrying out detection model training and classification model training; S5, the positions and the number of the vehicle-mounted articles are determined; S6, accurately identifying the article; and S7, identifying, recording and displaying the vehicle information. According to the invention, warehouse-in and warehouse-out conveying images can be recorded in real time, traceless searching in the subsequent storage inventory processis avoided, the whole process does not depend on participation of people, the types and the number of warehouse-in and warehouse-out articles are recorded in real time by accurately detecting and identifying the articles in the video images, and high efficiency, real-time performance, accuracy and intelligence of warehouse warehouse-in and warehouse-out article supervision are ensured.

Description

technical field [0001] The invention belongs to the technical field of deep learning, and in particular relates to a deep learning-based video surveillance warehouse entry and exit counting system and a deep learning-based video surveillance warehouse entry and exit counting method. Background technique [0002] For the physical product industry, warehouse management is an important part of enterprise management. Due to the variety of products, frequent in and out of the warehouse, a large amount of data needs to be processed quickly and accurately every day, otherwise it will affect the business development of the enterprise. Therefore, warehouse management has always been An important and daunting task for enterprises. [0003] With the introduction and popularization of network and computer technology and deep learning, the rapid development and implementation of artificial intelligence has led to the rapid development and implementation of artificial intelligence, making...

AI Technical Summary

Problems solved by technology

[0004] The traditional warehouse entry and exit supervision method usually relies on people to perform accounting, record the type of product, the time of quantity entry and exit, and then enter it into the warehouse management system for final warehouse inventory and proofreading, and for raw materials and processing plants, etc. Warehouse, raw materials are frequently transported, the time for warehouse items to enter and leave the warehouse is arbitrary, and the requirements for people are more stringent. It needs to be counted and calculated on call. It cannot be carried out in real time, and it will consume a lot of manpower and material resources.

[0005] In recent years, some people have also proposed a solution for inbound and outbound management, which is to use infrared code scanning guns to record items, which can distinguish item categories and record quantities. However, this method requires affixing a QR code on the item in advance and also It requires people to operate on the spot, which increases the cost and cannot completely avoid human participation; therefore, it is of great significance to provide a video surveillance warehouse entry and exit counting system and method based on deep learning for the above problems

Images