Information Fusion System Of Distributed Image Sensor Based On Intelligent Traffic

With the rapid development of urbanization and the increase of vehicle ownership rate,the traffic pressure is increasing.Therefore,how to monitor and deal with the problems in traffic road by intelligent means becomes the main research direction of intelligent transportation in the transportation system.In the intelligent transportation system,computer vision system plays an increasingly important role,through a variety of image sensors and computer vision processing system to achieve global road monitoring possible.In this paper,a panoramic camera is proposed as a sensing device in the intelligent transportation system to replace the current gun camera.It can not only reduce the density of the existing camera,but also expand the monitoring range.Using the camera to position the information,Exchange and tracking,to achieve intelligent traffic distributed image sensor information fusion and global monitoring.The main works are as follows: In addition,based on the images captured by the panoramic camera,the object is tracked with the Haar-like feature as the target description feature,and the software is used to compute the features in the panorama circle map,improved tracking effect on map transformation.Finally,a multi-camera tracking experiment is carried out by combining the target position information and the color histogram feature with the mean shift method.Firstly,based on the characteristics of the reentrant panoramic camera,an improved method of pixel mapping transformation is proposed based on the conventional image mapping deployment technique.The transformed image has the visual effect of perspective transformation to meet the requirements of the transformation of the image can better adapt to the target tracking.Based on the need of distributed imaging system to monitor target information fusion,the paper focuses on the calibration method of camera in world coordinate system in distributed system,and determines the world coordinates of each camera with the common visual target of multi-camera.The entire surveillance system is associated with all cameras.Once the monitor interest target of any location,once the monitoring system,distributed sensor arrangement through the exchange of information,can perceive its position,so that the control center can use its location information for target tracking and monitoring to provide the best sensor.Taking into account the imaging characteristics of the panoramic camera,the method of determining the image parameters of the blind ring is studied by using the circular blind location fixed by the image center,and the blind area is realized by image filtering,image binarization and Hough transform.Calculation of Center Darkness Center and Radius of Circle Graph.In addition,based on the images captured by the panoramic camera,the Haar-like feature is used as the target description feature,and the tracking algorithm is used to track the target.By compiling the software,the feature is compared with the panoramic view,and the improved mapping transformation.Finally,the multi-camera tracking experiments are carried out by means of mean shift based on the target position information and color histogram features.