Almost every CCTV camera we use today for security purposes only producing a video feed of particular area. Thus, manually analyzing and trigger events based on the real-time video feed is obviously a very difficult task. By this project REALTIME INTELLIGENT SURVEILLANCE FOR CCTV CAMERAS It is intended to do intelligent things like detect people, track different person separately and trigger events on business and security level based on detections from the ordinary video feed you get from the CCTV camera.

Technologies: Python,OpenCV,Dlib,Caffe neural framework and Flask

Functionalities of the system can be stated as follows.

• Detect humans in the video feed.

• Track each detected individual separately while they are in the video feed(without losing trackers at occlusions).

• Calibrating each tracker when its accuracy is less than a predefined threshold is done by each frame.

• Count people in a region at a given instance.

• Select and draw regions in the video feed and detect entrance to those regions.

• Detect people are walking or staying at particular place.

• Detect walking direction if they are walking.