Abstract: Wild animal intrusion has always been a persisting problem. A lethal conflict is below way among India’s developing population and its wildlife limited to ever-shrinking forests and grasslands. In forest and agricultural zones, human animal conflict is quite an issue where enormous amounts of resources are lost and human life is threatened. The reason behind animals attacking humans cannot be confined to a single cause. Certain animal attacks happen due to humans provoking them and others are purely based on instinct which is often the case and for which nothing can be done. There are no specific reasons for animals attacking humans based on instincts. In any way, animal attacks are daunting.

Apart from posing a threat to human life, Crop damage caused by animal attacks resulting in reducing the crop yield is also yet another consequence. Hence their activity must be monitored continuously in order to take action in case of animal intrusion in attack prone areas. Due to the diverse nature of movement and physical sizes of wild animals, it is a challenging task to track these animals or perform surveillance. In order to tackle the issue, we are developing a system to monitor these areas that will detect the intrusion of wild animals using image processing where classification is performed using Deep learning algorithms. Suitable action is taken based on the type of intruder and an alert is sent if the type matches the predefined wild animal datasets.

Keywords: Intrusion alert, Faster Regional Convolutional Neural Networks, Attack prone areas, Wildlife Datasets, Deep learning

PDF | DOI: 10.17148/IJARCCE.2023.124186

Open chat
Chat with IJARCCE