Outdoor Advertising Effectiveness Evaluation System Based On Intelligent Video Analysis

In recent years,the outdoor advertising media,represented by LCD and LED digital signage,is developing rapidly and the business scale is growing.As businesses invest more in advertising funds,the evaluation of advertising effectiveness has also become an important part of the whole product sales.At the present stage,the evaluation of advertising effectiveness is mainly based on artificial observation and questionnaire survey,and the means are backward.There are limitations in time and sample sampling,which can not be dynamically monitored during the advertisement.With the rapid development of video surveillance technology,intelligent processing of video information has become a cutting-edge research topic.Through video monitoring technology to achieve automatic evaluation of advertising effectiveness,it will have wide application significance.To solve this problem,this paper designs an advertising effectiveness evaluation system based on intelligent video analysis technology.By collecting the crowd scene video in front of the billboard,the system can automatically analyze the state of the pedestrian movement and the posture information,and finally make a scientific evaluation of the advertising effectiveness.First,pedestrian detection and tracking technology is used to count the number of pedestrians,then face detection is performed in the video to detect pedestrians who is facing the billboard.In order to detect whether pedestrians' sight is in the advertising area or not,a face orientation recognition method based on feature fusion and random forest algorithm is proposed to estimate face orientation accurately.For pedestrians who watch advertisements,the system automatically records the number of pedestrians,the watching time and the watching mode.Finally,the system uses feedback information to analyze and evaluate the advertising effectiveness.The experimental results show that the proposed algorithm based on feature fusion and random forest algorithm has achieved 97.16% recognition accuracy of face orientation on the face pose database,which is higher than other single feature classification methods.At the same time,compared with the traditional artificial evaluation method,the outdoor advertising evaluation system based on intelligent video analysis has a statistical accuracy of 88.8% for watching proportion and 96.1% for watching time proportion.It can effectively achieve pedes-trian information acquisition and classification,and make a scientific and reasonable evaluation of advertising effectiveness.