Intelligent Surveillance and Facial Recognition System for Efficient Border Monitoring and Threats Prediction Using Machine Learning Approach

Border authenticate an integral portion in a country's security, but borders are extremely prone and vulnerable to unauthorized migration, illegal smuggling of goods and drugs, human trafficking and assaults by the extremists as well as intrusion and cohesion between several parties. Border surveillance is the most central task in the area of national security and defense. Basically, borders need to be secured and kept under close monitoring and control for effective surveillance at different border points, which are stretched across hundreds of miles and have extreme terrain as well as climatic conditions. Hence, the need of the moment is to develop an automated border surveillance system which can perform the surveillance task without requiring any human intervention. The proposed system uses Internet Protocol (IP) camera or thermal imaging (TI) camera for detection of various objects and infiltrators. The Camera is assigned an IP address and connected through local network to the control center. Software code captures video and subsequently the intrusion detection. Machine Learning (ML) approach with Principal Component Analysis (PCA) and Viola Jones Algorithm was adopted for the intelligence recognition and facial detection of persons from list of images captured based on the surveillance IP camera. Consequently, the data sets were segmented into training set and test set of 300 images and 100 images respectively. Furthermore, we subjected our PCA model to an accuracy test and selected out different individual faces. These faces were all tested with our face recognition system and data was further computed in confusion matrix for performance evaluation and system accuracy. The confusion matrix demonstrated an overall accuracy of 83%, which indicates a good outcome of the Viola Jones Algorithm in our intelligent border surveillance system (IBSS).