多摄像机下运动目标跟踪识别系统

onsidering the limited vision field of the single-camera, in the monitoring of a wide range of scenarios, such as the college campus, shopping malls, the non-overlapping multi-camera system is in request .How to tracking targets in the non-overlapping multi-camera surveillance system is still an important problem because in the multi-camera system targets are separated in time and space. This paper presents a novel Two-Color histogram and LBP algorithm, which extract texture informations in the matching of the moving target. Also it use K-means Clustering algorithm and GMM algorithm in the multi-camera topology estimation. The system has a better detection reusult in the multi-camera target detection.