Community safety protection method and system based on multi-source heterogeneous data fusion

Abstract

The invention relates to the technical field of community safety protection, and particularly discloses a community safety protection method and system based on multi-source heterogeneous data fusion.The method comprises the following steps of acquiring an image of a surveillance camera; respectively inputting the image into a target tracking model, a behavior analysis model and a human head detection model; obtaining the location of a target, and meanwhile, carrying out behavior analysis to obtain behavior information, counting the number of human heads in the image, and obtaining coordinates of each target; carrying out target tracking, carrying out early warning on abnormal behaviors, and computing a human head distance; counting the number of people when the human head distance is smaller than a distance threshold; carrying out gathering early warning when the counted number of people exceeds a threshold value of the number of people; fusing passing records, pedestrian detection information and target tracking of infrared sensors, and comprehensively judging illegal invasion; and fusing results of target tracking and behavior detection, and carrying out key monitoring.

Description

technical field [0001] The invention relates to the technical field of community security, in particular to a community security protection method and system based on fusion of multi-source heterogeneous data. Background technique [0002] Community security protection is the most noteworthy module in both traditional communities and smart communities. Whether a community is good or bad, safety has always been an important reference point. The development of multi-source heterogeneous data fusion and deep learning can solve many problems in the automatic security protection of smart communities. [0003] Convolutional neural network is a method of deep learning. It uses its excellent performance in image processing and other fields combined with traditional algorithms to greatly reduce manual intervention and train high-quality models to improve accuracy. [0004] Multi-source heterogeneous data fusion was first used in the military field, but its ideas can be used to solv...

AI Technical Summary

Problems solved by technology

[0005] At present, in terms of community security protection, traditional monitoring equipment, sensors, and manual security inspections are mostly used. The data of each device is not effectively combined, forming a data island, which cannot be automatically formed into a security protection system. The labor cost is high and cannot be discovered in time. Occurrence of community safety incidents

Images