A Comparative Study of VGG19, Dense Net and Inception V3 for Wildlife and Human Threat Detection

The invasion of wildlife poses serious threats to farmlands since they destroy crops which results in losses being incurred by farmers. Fencing as well as manual surveillance are time-consuming and incredibly expensive. This particular research looks into how artificial intelligence can help farmers by using deep learning algorithms for recognizing animals and humans in farm peripheries. It is true that AI has made major advancements, but there are very scarce studies that focus on comparing multiple models to check which has the best accuracy, speed, and adaptability. The researchers analysed a few of the most renowned models and deep model architectures like VGG19, DenseNet, and InceptionV3 on datasets that focus on human and wildlife threat detection. This research also tries to show that InceptionV3 performs the best and is the most accurate while most reliable for detection tasks, and that DenseNet performs well in balancing accuracy to speed. Analysis performed in this research helps narrow down the AI system best suited for farm security, which gives insights into improving security AI developed systems for farms. Crop security solutions are easier and faster to develop for farmers, aid developers and researchers, which positions this work as a pioneer. Wildlife detection through AI reduces the need for supervision from farmers, which makes the process safer and more effective in protecting farmlands. Developing reliable and cost-effective ways to monitor farmland security while minimizing crop damage can be done by this research. This establishes a foundation for future work around the field as it focuses on AI based project aimed at farm security.