Method for improving credibility of cross-camera behavior recognition by using block chain

Abstract

The invention discloses a method for improving the credibility of cross-camera behavior recognition by using a block chain, and the invention comprises a camera module, a block chain module and a behavior recognition module, and the method comprises the following steps: 1, defining a single-chain block chain service; 2, defining a camera Internet of Things system capable of automatically segmenting a video, calculating hash as a file name for storage, packaging and transacting the file hash, signing and sending the file hash to a block chain storage evidence; Step 3, enabling the camera service to poll the block height regularly; segmenting the video according to the block outlet time of each block of the block chain, wherein the blocks are associated through Hash, data in the blocks of the single chain are progressively increased according to the sequence of block heights, and the video with the index evidence is progressively increased according to the block outlet sequence; and meanwhile, determining the event occurrence time according to the block timestamps to ensure that the event sequence is not disturbed. According to the invention, the credibility of a cross-camera behavior recognition result can be improved, and the possibility that the behavior recognition result is doubted is reduced.

Description

technical field [0001] The invention relates to the field of behavior recognition credibility, in particular to a method for improving cross-camera behavior recognition credibility by using blockchain. Background technique [0002] Compared with algorithms that let machines learn, machine learning is about letting machines learn. The earliest application of machine learning is to distinguish spam. The traditional way of solving problems by computers is to write rules, define "spam" and let computers execute them. This requires writing an algorithm. This algorithm has a fixed input and output. The input is a piece of email corresponding to all the information, and the output is to judge whether the email is spam. [0003] At present, the algorithm for identifying human behavior through machine learning has become mature, but machine learning is to calculate the output corresponding to the input of pictures or videos, and it cannot judge the sequence of events. Therefore, the...

AI Technical Summary

Problems solved by technology

[0003] At present, the algorithm for identifying human behavior through machine learning has become mature, but machine learning is to calculate the output corresponding to the input of pictures or videos, and it cannot judge the sequence of events. Therefore, the behavior recognition with event sequence judgment , it is difficult to ensure the accuracy of machine learning results, so this invention proposes a method of using blockchain to improve the credibility of cross-camera behavior recognition to solve the above problems

[0004] For example, application number 201811353903.2 discloses a video active recognition method combined with blockchain, which is applied in the field of video recognition, in order to completely and truly save video frames and prevent tampering; , encrypt and save the video file and verify the source of the video and whether it has been tampered with. The private blockchain in the present invention can be used as a local database to provide a basis for verifying the source of the video and whether the video has been tampered with; and the first A transaction timestamp can be used as the time of occurrence of abnormal events to ensure the immediacy of the time of locally stored video content, but the invention itself cannot accurately determine the sequence of events occurring

Images