Dangerous behaviors detection based on deep learning

Abstract

Deep learning has a high degree of popularity in recent years. It is widely used in computer vision, artificial intelligence and other fields. Sites with high safety needs, such as gas stations, have a high demand for monitoring of dangerous behaviors such as smoking. Under normal circumstances, gas stations will employ corresponding personnel to inspect and supervise, but such labor costs are higher, and the monitoring effect is not good. This article is to use an object detection system based on deep learning technology to detect the dangerous behavior of gas stations. This article mainly solves several problems for gas stations to detect dangerous behaviors: first, what technology is used to achieve object detection; secondly, how to increase the speed of detection as much as possible; and thirdly, how to improve the accuracy of detecting dangerous behavior. To solve the above problems, this article will introduce how to implement an object detection system based on deep learning technology. First, a data set containing dangerous goods is established, then the convolutional neural network is trained, and finally the test results of the training results are checked and transplanted. The results prove that the gas station dangerous behavior detection system based on deep learning technology realized can accurately and quickly detect dangerous objects (cigarettes, etc.) in the image.