Animal Activity Recognition Using Convolutional Neural Network
R. Subraja, Y. Varthamanan · 2023
It will be beneficial to observe and track animal behavior in order to avoid human-wildlife conflicts HWC. The introduction of wild animals produces multiple losses in a variety of biome zones, with agricultural loss accounting for a large portion of the losses. This conflict, in addition to crop loss, destroys human habitat and claims lives and property. Because loss occurs on both sides, this research may provide a solution to stop the loss of human and animal lives. In the proposed work, we use a mechanism to detect the animal correctly and sensitively. When an animal is discovered, the warning system emits a loud beep, alerting forest officers and those nearby to take appropriate action. The animal was recognized in this scenario utilizing a convolutional neural network filled with GLCM features and the deep learning method. The obtained image of the animal goes through several steps before being trained and checked against the database we collected. Finally, 98% of the desired production is obtained.