A weighbridge human intrusion detection method based on mask Rcnn and SSD

Abstract

The invention discloses a weighbridge human intrusion detection method based on Mask Rcnn and SSD convolutional neural network. The historical video of weighbridge area is collected by a camera; when the weighbridge weighing state is an idle state, the Mask Rcnn convolutional neural network is used to identify the real-time weighbridge area. The weighbridge area of ​​the video and save it; the pedestrian vehicle model trained by the SSD convolutional neural network recognizes the position information of the person and the vehicle in the real-time weighing video. When the position information of the person intersects with the weighbridge area, it means that there is a human body in the weighbridge area. Intrusion, record the weighing state, the number of people in the surrounding environment, and the number of people in the weighbridge area; record the human intrusion time; when the human intrusion time exceeds the set threshold, it will automatically alarm and record the abnormal weighing information. The invention can carry out real-time monitoring on the on-site vehicle weighing process, alarm the human intrusion behavior, and can effectively identify the cheating phenomenon generated when the truck is weighed.

Description

technical field [0001] The invention belongs to real-time image processing technology, in particular to a weighbridge human intrusion detection method based on Mask Rcnn and SSD. Background technique [0002] As an accurate, convenient and fast weighing and measuring equipment, electronic truck scales have been increasingly applied to various industries such as logistics, steel, coal, electricity, building materials and so on. However, along with the pursuit of interests, there is also the phenomenon of using electronic truck scales to implement measurement cheating in the weighing process, which has brought significant economic losses to enterprises and customers, and also brought challenges to the integrity of today's society. Eliminating these cheating phenomena is a problem that needs to be solved urgently. [0003] In the weighing and weighing process, there are the following cheating behaviors: during the weighing process of the truck, someone illegally breaks into th...

AI Technical Summary

Problems solved by technology

Infrared image detection uses a Gaussian model for image segmentation, and then extracts histogram of oriented gradients (HOG) features, and uses SVM or Adaboost for classification and recognition. In the case of insufficient illumination and low contrast, infrared radiation imaging has a good detection effect, but Infrared detection design and maintenance costs are high

However, the problem of traditional image detection technology is that the scene image background is relatively complex, and the surrounding environment will interfere with human intrusion detection. The success rate of recognition for the weighbridge area of ​​different angles and different sizes of people and vehicles is low, and the image processing time is slow. Cannot meet the requirements of real-time detection

Images