Wildlife Exploration App

Computer Engineering

This initiative addresses the pressing need for monitoring wildlife in real-time within border regions of jungle areas and forest reserves to mitigate potential human-wildlife interactions and enhance conservation efforts. A Python-based system has been developed that integrates computer vision and instant messaging technologies for the autonomous identification of animals and notification. The core of the system utilizes the pre-trained YOLOv5 object detection model, which has been trained on an extensive dataset, to identify specific animal species within its viewing area. To optimize detection for its particular use case, the model's confidence threshold has been adjusted to 10%, thus enhancing its sensitivity. The system is designed to identify a specified set of animal classes relevant to the unique ecological context, as articulated in a YAML configuration file, allowing for adaptability across different geographical locations and conservation goals. A regular laptop camera captures live video, and the YOLOv5 model processes each frame to identify objects. The initial detection results are refined using a custom filtering function (filter_animals) to locate the specified animal categories. Upon identifying a target animal, the system promptly sends an alert via WhatsApp, employing the pywhatkit library to deliver a message to a designated recipient, such as village officials or forest rangers. This alert message specifies the type of animal that was detected. Additionally, the system provides visual feedback by showing bounding boxes around the identified animals on the video stream, which is presented in real-time using cv2.

This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.