Sandalwood is one of the most valuable trees in the world. Sandalwood robberies are increasing day by day as there isn't enough security being provided for them. Eventually they are getting one of the most endangered trees in our country. So, to protect the trees, in this paper proposed a method using inception-v3 model. The main objective is to detect if there is any person in the forest particularly in the restricted time. The importance of infrared images are that these can be visible even in the night time (i.e. even if no light). The main role is to create a classifier which detects the type of weapon hold by the person. This paper is based on Inception-v3 model of tensorflow. We use the technique transfer learning to retrain the infrared images. The implementation is done based on benchmark dataset OTCBVS under 'Terravic Weapon Infrared Database'. This paper propose a model with the 99% accuracy. © 2018 IEEE.