Intelligent video analyzing method for pedestrians crossing zebra stripes

Abstract

The invention discloses an intelligent video analyzing method for pedestrians crossing zebra stripes. The intelligent video analyzing method comprises the following steps that (1) a video analyzing system is arranged near the zebra stripes and is started regularly at preset time points, and cameras collect video information data in real time and transmit the video information data to a video analyzer; (2) the video analyzer analyzes the video information data and judges states of traffic lights, the intruder alarming function of an intelligent alarming module is started if the red light is turned on, and the prompt information playing function of the intelligent alarming module is started if the green light or the yellow light is turned on; (3) the video analyzing system is closed at preset time points, and the video information data of the day are stored before the video analyzing system is closed. By means of the scheme, the unique intelligent video analyzing method for the urban zebra stripe traffic lights and the moving pedestrians is achieved, the intrusion on the pedestrians crossing the zebra stripes is detected when the red light is turned on, and intelligent audio and video alarming is conducted according to detecting results.

Description

technical field [0001] The invention relates to the technical field of video analysis methods, in particular to a video analysis method for intelligently analyzing pedestrians crossing a zebra crossing and running a red light. Background technique [0002] At present, the urban population is increasing, and the number of vehicles driving in the city is increasing at an explosive rate. Due to the fast pace of urban life, there are always scenes of life and death on the zebra crossing, and a Chinese style of running a red light is formed. Get used to crossing the road. [0003] In order for pedestrians to regard walking the zebra crossing as a fixed custom, and not crossing the zebra crossing with red lights to become a virtue, it is necessary to publicize and publicize civilized traffic to the general public, and how to let the public have a good way to get publicized And the effect of on-site supervision, generally speaking, installing a civilized traffic publicity device n...

AI Technical Summary

Problems solved by technology

Although this invention can capture the actions of pedestrians crossing the road in real time and use this to judge whether the pedestrian has run a red light while crossing the road, and analyze and monitor human body movements in real time, but to realize the real-time analysis process, it is necessary to first establish a three-dimensional human body Model, by obtaining static human body posture image sequences at different sampling moments to track and calibrate the human body movement state, and matching the image with a predefined human body model, so as to obtain the human body movement parameters at each sampling moment. If the number of static human body posture image sequences is too small, there will be a problem of low accuracy, and if in order to improve the accuracy rate, it is necessary to collect a large number of static human body posture image sequences at different times, and pedestrians crossing the zebra crossing at every moment of every day have different are not the same, so even if the number of collections is large enough, it still cannot achieve high accuracy, nor can it perform adaptive updates according to new changes in pedestrians, so its accuracy cannot meet a very high requirement; secondly, in the process, It is necessary to match the collected images with the existing human body models. It is necessary to first perform image processing and local matting on the collected images, and then match them with a large number of models in the database. The calculation of the whole process is complex, and the hardware High requirements and long calculation time are not suitable for instant alarm processing at the zebra crossing position

Images