Intelligent Farm Surveillance System For Bird Detection

Abstract

Intelligent farm surveillance system refers to video level processing techniques for identi cation of objects from farms video. Many developed and developing country are using intelligent farm surveillance system for viewing the farm remotely. Here in this thesis we are having one of the video from farm surveillance system and from that we are detecting the bird and then ringing the alarm. This can be useful for protecting the farm from crop hazard by birds. In this thesis there is a brief survey of di erent object detection techniques including many background subtraction techniques like frame di erencing, sum of absolute di erences, single and mixture of Gaussian model etc. Further for identifying object as a bird there are di erent techniques like tem- plate matching, skeleton extraction, contour based technique, edge based technique, etc. Here after implementing di erent methods and combining best feature of them, the system is proposed for bird detection. Proposed system uses the combination of frame di erencing and gaussian mixture model for background subtraction. It uses normalized cross correlation method for template matching.Proposed system is hav- ing relatively low false positive rate and false negative rate.