Embedded human detection system based on thermal and infrared sensors for anti-poaching application

摘要

Poaching for massive traditional medicine demands is the single greatest threat of extinction that looms over most wildlife especially the endangered tiger. One way to curb the poaching activity is to have tighter monitoring of the protected area by using camera traps. In this research, an embedded system with human detection features for anti-poaching application is proposed. The system uses a 4 × 4 pixels' thermal sensor to detect any heat presence before an infrared camera is activated to capture video images. The images will be processed with a human detection algorithm based on frame differencing and orientation features. The detected human will then be verified using a Support Vector Machine (SVM) classifier. Once a human is identified, the image and the location of the human can then be sent through a long range data connection to a ranger office to alert the authorities. The system was implemented on a Raspberry Pi 2 board. Experiments have been conducted to evaluate the functionality of the system in non-urban environment during day and night times. The results showed that it can successfully detect human in an average of less than 2s even during night time with complete darkness.