Paper Title
WEED MANAGEMENT BY CONVOLUTIONAL NEURAL NETWORK

Abstract
In this research, we present a novel approach for precision weed management in agriculture by integrating drone technology with Convolutional Neural Networks (CNNs). The methodology involves the use of drones equipped with high-resolution cameras and Global Positioning System (GPS) modules to capture detailed images of agricultural fields. These images are preprocessed and used to train a CNN model capable of accurately detecting and classifying various weed species. The trained model generates weed density maps, which are utilised to create prescription maps for targeted herbicide application. Our results demonstrate a high weed detection accuracy, with a precision of 92% and a recall of 89%. The use of prescription maps resulted in a 30% reduction in herbicide usage compared to traditional methods. This approach not only enhances weed control efficiency but also promotes sustainable agricultural practices by minimising chemical usage and preserving non-target plants. The system's decision support capabilities provide real-time alerts and recommendations to farmers, further improving operational efficiency and reducing labor costs. Overall, this research highlights the potential of integrating advanced drone and machine learning technologies to transform weed management practices in modern agriculture. Keyword - Convolutional neural network, kernel, Ground sample distance, Visual geometry group (VGG), Residual network (ResNet), RGB Camera, Hyper-spectral imaging.