Bina içi akilli görüntü güvenlik sistemi uygulamasi

In this study, an application system of indoor intelligent distributed video surveillance designed and developed with present image processing algorithms and application programming interfaces. This study aims to develop a supplementary video surveillance system on a pre-installed passive video surveillance system. The passive video surveillance systems record surveillance video feed with movement detection or plain full time interface. In this study, video feeds recorded by movement detection, will be analyzed for pedestrian and face extraction. After pedestrian and face data extracted from video feeds, pedestrian data will be grouped for every entering and leaving moments with matching process, then each image group of pedestrian's faces will be analyzed to find reappearance of the same pedestrian. After all image processing, system will predict each unique pedestrian's every entering and leaving time. This study used OpenCV as primary API to manage grabbing, processing and post-processing of video and images. For application software architecture, Qt application programming framework utilized on thread and GUI programming. And for gathered information and statistical purposes SQLite database used. In this study, for human detection, Histogram of Oriented Gradients and face detection algorithms tested to find out optimum algorithm for human counting and recognition on videos of indoor video surveillance.