Classification of Aerial Cactus for Conserving Biodiversity Hotspot Zones using Deep Convolutional Neural Network VGG16

There are many exceptional regions in the world with the existence of varieties of unique several endemic plants in biodiversity. Conservation of biodiversity 'hotspots' is protection and maintenance in a sustainable way. The research studies proved a good relationship with a diverse ecosystem would help health, resource consumption, climate changes, and other areas positively. Human activities which lead to the destruction of biodiversity. So our aim is to identify this biodiversity-ecosystem by using automated surveillance and preserve the ecozone. To assess the impact of earth's natural resources, it is necessary to build a model that identifies the columnar cactus in the aerial image and recognizes the vegetation inside the protected areas by identifying the columnar cactus. The proposed model aims to identify and to classify aerial cactus with the help of VGG 16 Convolutional Neural Network architectures. The main work is done with the help of experimentation and evaluation of the performance of the network to recognize a specific type of cactus in aerial imagery. The dataset containing the 21,500 images, this images is provided by the Kaggle and it is a part of the VIGIA project. The proposed model compared with various convolutional neural network models like ResNet26, hyperspectral CNN, LetNet-5 and Resnet50 for the particular problem. The experiment result of the proposed model shown the high prediction accuracy of 98% and 0.98 ROC/AUC score, substantiate the purpose of the perspective to accomplish best in results for the columnar cactus recognition.