Research On Indoor Occupant Counting Methods Based On Zenithal Video

Occupant information is an important attribute in the building space.The accurate acquisition of this information is very important to the intelligent operation decision of building equipment and the energy saving decision of building operation in intelligent buildings.The occupant counting method based on video analysis technology is concerned by academia and industry because of its advantages of low cost,simple installation and high precision.In order to meet the demand of new building intelligent system platform for the number of occupants in buildings,the indoor occupant counting method based on zenithal video is studied.In order to solve the problem of camera installation height has influence on the accuracy of indoor occupant counting method and the problem that indoor occupant counting has accumulated error,the indoor occupant counting method is improved.Finally,a set of indoor occupants counting system based on zenithal video is designed and implemented.First of all,aiming at the video shot from the top view of the entrance and exit has the characteristics of simple background and less occlusion between objects,this paper proposes a method of indoor occupant counting based on the zenithal video.The method uses KNN(K-Nearest Neighbor)algorithm to detect the moving targets in the verhead surveillance video of the entrance and exit area.The Kalman Filter(KF)algorithm is used to track the detected moving targets.According to the moving track of the moving target and the relative position of the boundary line,the direction of the moving target is judged.The identification of occupant target is realized according to the movement characteristics of occupant target.The number of occupants in the room is obtained by counting the number of occupants entering and leaving the boundary.The experiment shows that the accuracy of the method can reach more than 90% in terms of the number of occupants in the room,the accuracy of indoor occupied or unoccupied state and the accuracy of the staying time of indoor occupant.And the average processing rate of each frame can meet the demand of online video analysis.Secondly,considering that different installation heights of cameras in practical application will affect the size of detection area and the setting of the parameter values of human movement characteristics in the occupant counting method,and then affect the accuracy of the occupant counting method.Aiming at the problem,an automatic method to obtain the installation height of the camera is designed.And according to the relationship between the camera installation height and the object imaging length,the object motion speed,the key parameter configuration table in the occupant counting method under different heights is given.The camera installation height can be automatically obtained and according to the height the optimal parameters can be selected,eliminating the influence of different camera installation height on the accuracy of the algorithm.Aiming at the problem of cumulative error in the occupant counting method in this paper,a correction method of indoor occupant number based on the occupant base is designed,and the effectiveness of the correction method is proved by experiments.Finally,in order to meet the demand of the new building intelligent system platform for the perception of the population distribution in the building area.The indoor occupant counting method based on zenithal video is used to design and implement the indoor occupant counting system based on zenithal video.The system is composed of a webcam,a local information processing node and a CPN(Computing Process Node),which is connected wirelessly to realize the transmission of video data between the webcam and the local information processing node.The webcam is responsible for collecting video data,and the local information processing node is responsible for analyzing and processing the collected video data.And the occupant information related to the entry and exit of border and the number of occupants in the room are stored to the local database server.The demonstration application shows that the system can satisfy the data support when the new intelligent building system platform dispatches the intelligent building equipment and makes the energy saving strategy.Figure [45] Table [26] Reference [55]...