Application Of City Surveillance System Quality Evaluation Method

With the intelligent city construction,video surveillance system has been utilized in many scenes.Until now,the number of cameras in each city has reached tens of thousands,which means that testing the quality of the video surveillance system entirely manually is impractical.Therefore,an effective quality evaluation method for the surveillance system is necessary.Some studies find that people usually require the high video quality,effective monitoring area,and the target validity in the video surveillance system.Based on these three requirements,this thesis proposes amore comprehensive quality evaluation methodfor intelligent video surveillance system.Firstly,in terms of three detection modules,i.e.,the video quality,the effective monitoring area,and the target validity,we respectively establish the subjective and objective evaluation.Then,we use three detection algorithms to detect one monitoring camera and calculate thesubjective score.The subjective score of surveillance system is the weighted sum of three subjective scores.Finally,the evaluation score of whole surveillance system is obtained by summing the score of all cameras with different weights according the importance of monitoring areas.In this thesis,the subjective and objective experiments are conducted by selecting several testing videos.This experiment includes the evaluation of video quality,the effective monitoring area and the target validity.The analysis of experimental results illustrates the effectiveness of the proposed subjective and objective evaluation approaches.The detection results of proposed approach arebasically consistent with the manual evaluation result.But with different weights of detection modules,the results of very few of videos differ slightly.In the future work,we need to design an optimization strategy for weights selection,in order to make the subjective and objective results more consistent.