Research On Indoor Occupant Counting Method For Insect Intelligent Building Management Platform

In the research and practice of smart buildings and building energy efficiency,the number of occupant,the location of occupant,and the presence or absence of occupant are important attributes in the building space.This paper designs a two-stage indoor occupant counting method based on cascade classifiers and convolutional neural networks facing the acquisition of occupant distribution information in the building space.This paper also discusses the influence of the correct rate of occupant counting and the accuracy of identifying the presence or absence of occupant in the indoor static background and the dynamic background.And the indoor occupant counting method was improved based on the partial elimination of the static background and the elimination of the dynamic background.Finally,a set of indoor occupant counting system based on image analysis is designed and implemented for the insect intelligent building management platform.First of all,this paper designs a two-stage indoor occupant counting method using cascade classifiers and convolutional neural networks based on a single camera.This method uses the Adaboost cascade classifier to screen out candidate targets,and uses the CNN classifier to accurately identify the candidate targets and obtain the number of occupant.Furthermore,in order to solve the problem of blind spots in a single camera,this paper designs a method for checking the number of indoor occupant based on the area covered by the two cameras installed diagonally at the same height.Secondly,In order to eliminate the influence of background on human target detection in building space,this paper proposes an improved algorithm for indoor occupant counting based on background elimination.Aiming at the scene with relatively fixed background,this paper proposes an improvement method based on the partial elimination of static background.Experiments show that this scheme can not only improve the detection accuracy,but also improve the detection efficiency when selecting a suitable background elimination coefficient.Furthermore,this paper proposes an improved scheme based on dynamic background elimination in order to adapt to the constantly changing scenes.This scheme is based on Gaussian mixture model and background subtraction to achieve dynamic background acquisition and elimination.Experiments show that this method can effectively improve the accuracy of occupant counting and the accuracy of identifying the presence or absence of occupant when the background changes dynamically.Finally,an indoor occupant counting system was designed and implemented combined with the indoor occupant counting method and the improved scheme of background elimination facing the needs of the insect intelligent building management platform.The system consists of a web camera and a local information processing node,which are connected wirelessly.The web camera is responsible for collecting image data,and the local information processing node is responsible for analyzing and processing the image.The processing results can be directly uploaded to the CPN node in the insect intelligent building management platform through the local port.The application shows that the system can meet the needs of the insect intelligent building management platform for independent perception of the number of occupant in each area.Figure[45] Table[28] Reference[73]...