Method and system for pedestrian retrieval based on internet of things

Abstract

The invention, which is applied to the video monitoring field, discloses a method and a system for pedestrian retrieval based on the internet of things. The method comprises: obtaining video streams that are collected by cameras under an association scene, extracting candidate streak lines from the video streams by respectively utilizing an edge detection algorithm, and utilizing a horizontal line or a vertical line to carry out approaching on a candidate streak line meeting requirements; carrying out streak line density determination and removing a non-pedestrian interference so as to obtain a framed pedestrian position; extracting a streak clothes pedestrian inter-frame characteristic to establish a characteristic database; according to the front and back position relation of cameras under the association scene, selecting a to-be-retrieved streak clothes pedestrian in a current camera, extracting the position and the characteristic, and carrying matching and ordering on the extracted position and the characteristic and the characteristic database in a previous camera in a certain time range, so that a pedestrian retrieval result is obtained. According to the embodiment of the invention, a streak clothes pedestrian retrieval technology based on an internet of things concept and multi-camera information is employed; and compared with the prior art, the employed technology enables a streak clothes pedestrian to be identified rapidly and efficiently, thereby well meeting demands on traffic enforcement by the police.

Description

technical field [0001] The invention relates to the technical field of video monitoring, in particular to the field of intelligent security of the Internet of Things, and in particular to a method and system for pedestrian retrieval based on the Internet of Things. Background technique [0002] With the rapid development of multimedia technology, video data is becoming larger and larger. How to analyze, search and identify these massive video data has become an important challenge in related research fields, especially in the fields of intelligent monitoring and digital security. Hotspots and Difficulties. [0003] Pedestrian retrieval technology based on video image is a kind of recognition technology that uses image algorithm to analyze the characteristics of the acquired video stream, and finally retrieves pedestrians. Pedestrian retrieval technology based on video images has the advantages of wide detection range, large amount of acquired information, flexible installat...

AI Technical Summary

Problems solved by technology

[0004] The purpose of the embodiment of the present invention is to propose a method and system for pedestrian retrieval based on the Internet of Things to solve the problem that the traditional pedestrian retrieval technology based on video imagery relies too much on classifiers to distinguish pedestrian and non-pedestrian movement prospects and then perform feature matching. Based on the Internet of Things, the solution can directly use image sequence features to retrieve pedestrians in striped clothing, and can effectively eliminate non-pedestrian interference such as vehicles, and provide law enforcement and other needs in the criminal investigation field.

Images