Ecological Factors Observation and Animal Identification System

Traditional ecological observation and wildlife monitoring lacks comprehensive environmental sensing and thus are inefficient in offering insightful analysis. In recent years, integration of deep learning and robotic solutions have revolutionized various fields. In response to the limitations, this study introduces development of an innovative autonomous system for comprehensive wildlife conservation and environmental data collection. The system presents a robo car, incorporating a suite of sensors for data collection and the camera module, leveraging an advanced CNN model with ConvNeXt architecture, fine-tuned on 90 different animal species. Unlike traditional monitoring systems, the robo car navigates the forest area, allowing reach to diverse and previously inaccessible areas especially in challenging or dangerous terrains. Further, the use of ConvNeXt enables prompt responses and precise tracking of wildlife with an accuracy of about $97 \%$. This system not only advances the field of ecological observation but also offers valuable tools for biodiversity studies, forest conservation and environmental management, setting new standards for autonomous ecological monitoring solutions.