Early-warning detection method for emergencies in smart city video monitoring

Abstract

The invention relates to an early-warning detection method for emergencies in smart city video monitoring. The method comprises the steps that an early-warning detection system is established, and in order to ensure monitoring timeliness, a video is uploaded through double channels; a client preprocesses a monitoring video stream, and a local host only extracts foreground features in key frames through adoption of a background model and through key frame extraction, background separation and foreground feature extraction, the local host packages and compresses the foreground features and uploads the foreground features to a cloud server; moreover, the video stream is uploaded to the cloud server through another channel, a classifier is established through frame interception, object detection segmentation and object feature extraction and through adoption of an SACBA adaptive association rule classification algorithm, and through continuously increasing original video data expansion and through updating of an abnormal feature library, and the accuracy of detecting various emergencies is gradually improved; and the cloud server compares the received packaged feature data packet information with the abnormal feature library, carries out pre-warning judgment and feeds back a result to the client. According to the method, the early-warning monitoring timeliness and accuracy can be ensured.

Description

technical field [0001] The invention relates to an early warning detection method for emergencies in video surveillance of a smart city. Background technique [0002] There are many types of video surveillance, and traditional surveillance video anomaly detection methods cannot be well adapted to various environments. Because of the current detection method, each camera generates a huge amount of video every day. If it is processed locally, due to the diversity of video types and the limited computing power of the local host, the processing accuracy and response speed are very low when faced with massive surveillance videos. Low, so the detection and analysis on the local host, the accuracy of the early warning results cannot be guaranteed; if uploaded to the cloud server for processing, due to the huge amount of surveillance video data, the process of uploading to the cloud will take too long, and it will also be due to network congestion and other conditions Affecting the...

AI Technical Summary

Problems solved by technology

The disadvantage of this detection method is that the optical flow method is used to calculate the change of entropy value, and the calculation amount of the optical flow method is relatively large, which cannot guarantee real-time and practicality; Square video surveillance with square dancing, or crowded shopping malls, the motion energy of these scenes will cause a lot of false alarms to the monitoring system, and this method uses the same threshold for the change of entropy value, which is not practical

[0004] In the Chinese Patent No. 201210223375.5, the disclosed "Abnormal Behavior Detection Method for Group Crowds in Video Surveillance" uses a long-period motion estimation method based on video particles to obtain the trajectory of the target, and calculate the distance between the trajectories and the travel speed information. Spectral clustering analysis is used to identify blockages and falls through sudden changes in normal trajectories. However, the detection behavior of this method is too single. For abnormal scenes with inconspicuous motion features, there are a large number of false positives and negative negatives. The judgment work is concentrated on the local host. When processing multiple real-time video streams, the performance of the host computer is relatively limited

[0005] Chinese Patent Application No. 201610841596.7 "Method for Detection of Personnel Loitering Behavior Based on Video Surveillance Platform" discloses the use of Gaussian mixture background method for background modeling, and simple screening based on contour complexity, area and aspect ratio of circumscribed rectangles , to obtain the foreground block of pedestrians, and then further judge the degree of video state change, and then get the video stream of abnormal behavior. The judgment work is concentrated on the local host. When processing multiple real-time video streams, the performance of the host is relatively limited.

The disadvantage of this monitoring system is that the real-time video data of the whole city is huge, and the data transmission from the monitoring information collection end to the monitoring information processing end takes a long time, and it is easily restricted by the network environment, making it difficult to guarantee real-time performance

Images