Research And Implementation Of Indoor People Counting Method Based On Online Ada Boost Algorithm

Video surveillance equipment has been widely used in schools,shopping malls,banks,stations etc...Managing human density by detecting the number of people is meaningful in these occasion.The research and application based on the online AdaBoost algorithm for the people counting in indoor spaces Through the intelligent processing of the image and according to the pedestrian information in the video surveillance area count the number of people.It is the field of computer vision is very important a research direction,with a broad research prospects.In this paper,through the study of the background of the subject,the development of the status quo at home and abroad,has been widely used in the target pedestrian detection and statistical methods such as skin color matching,template matching,based on statistical methods were studied and compared.Paper get a favorable method and technology of people counting method in indoor spaces.The main work is as follows.1.This paper studies the basic principle and implementation process of human head detection based on online AdaBoost algorithm,and analyzes the calculation of eigenvalues in online AdaBoost algorithm and the online training process of classifier.Head detection and statistics use double iterative detection strategy to come out.2.A method of sample labeling is proposed to verify the results of the detection system by combining the head and shoulder edge characteristics.Determine the head foreground image,compare the verification of online AdaBoost algorithm test results are correct,the detection of the wrong area of the sample label,online AdaBoost classifier through the new sample to learn,update the classifier.3.On the basis of the head and shoulders verification system,we join the manual intervention system.Once the head and shoulder verification system is marked with the sample error,it will make the classifier classification accuracy is getting lower and lower,adding manual intervention,the false detection and leakage detection area marked as a new sample for online learning.Correct the detection performance of the system.4.Designed system complete testing and analyzing of function and performance.We test a large number of experiments based on classroom surveillance video data.The results show that our method is able to detect the number of people in video image accurately and quickly.Regardless of the change of crowd mobility and crowd density,this method still has better detection performance.The research work completes the target detection of the online AdaBoost algorithm,and proposes a sample annotation method based on the head and shoulders feature and the manual intervention sample annotation.To achieve the indoor people counting,through the experimental system to verify its effectiveness,accuracy and real-time.