This project is an attempt to prevent deforestation and animal poaching by building a Convolution Neural Network model. The model is trained on the graphs corresponding to the sounds of Chainsaw, Gunshots and the regular noise in forests (birds chirping, raining, sounds of differerent animals). The data used to train the model has been manually collected. The data size is about 7000 graphs, corresponding to each of the categories. The model gives a value 1 if it detects a Chainsaw sound, 2 if it detect a Gunshot sound and 0 otherwise. This repository only includes the software part i.e. the code used to train the model.