Time sequence and CNN-based unsafe behavior identification method and system

Abstract

The invention discloses a time sequence and CNN-based unsafe behavior identification method and system. The method comprises the steps of inputting original data of a training set video of unsafe behaviors to a CNN for performing training and learning, and outputting a spatial eigenvector from the last pooling layer of the CNN; taking the spatial eigenvector as an input of a time recurrent neuralnetwork, learning a time sequence dependency relationship contained in the spatial eigenvector by using the time recurrent neural network to learn a time sequence representation of spatial behaviors,and obtaining an average pooling eigenvector; inputting the average pooling eigenvector to a softmax layer, thereby obtaining a deep mixed learning model of the CNN and the time recurrent neural network, namely, a softmax classifier; and performing online identification on a monitoring video of a construction site by utilizing the softmax classifier to obtain the unsafe behaviors of the construction site. The support can be provided for real-time investigation and correction of the unsafe behaviors in a whole building engineering construction process.

Description

technical field [0001] The invention belongs to the field of construction engineering informatization, and more specifically relates to a method and system for identifying unsafe behaviors based on time series and convolutional neural networks. Background technique [0002] Due to the dynamics and complexity of the construction environment, ensuring the safety of construction workers during construction operations is a common and challenging problem. Despite the continuous improvement of laws and regulations, the continuous increase of supervision, and the unremitting efforts of industry practitioners to improve this problem, the number of accidents and deaths in the construction process is still high. According to Heinrich statistics, about 88% of accidents in construction are caused by unsafe behaviors. Therefore, in order to avoid serious casualties and property losses caused by safety accidents, domestic and foreign researchers have done a lot of research on the identif...

AI Technical Summary

Problems solved by technology

[0003] The initial identification of unsafe behaviors is based on manual observation, which can accurately identify unsafe behaviors in construction and propose corrective measures. However, due to complete reliance on manual labor, there are defects such as time-consuming, manpower-consuming, and subjectivity.

[0004] In order to solve the above defects, many researchers apply sensing technology (such as: RFID, UWB, GPS) to locate and track construction objects, so as to identify unsafe behaviors, but it is necessary to install detection equipment on each detection object, which will affect the construction to a certain extent. The normal operation of workers; moreover, the large number of equipment is not conducive to maintenance and management

[0005] Based on this, with the development of computer vision, researchers began to turn their attention to pattern recognition methods, but the process of identifying unsafe behaviors often relies too much on manual feature extraction, which involves cumbersome parameter adjustment processes; Insufficient, unable to reflect unsafe behavior dynamics in a timely manner

Images