Intelligent Insecticide and Fertilizer Recommendation System using CNN ResNet-50

Abstract— Smart farming through the development of anintelligent crop, insecticide and fertilizer recommendationsystem. Leveraging the power of Convolutional NeuralNetworks (CNNs) and the efficiency of ResNet50architecture, our system aims to optimize crop managementpractices in agricultural settings. The proposed systemutilizes Internet of Things (IoT) devices deployed acrossfarmland to collect real-time data on plant health,environmental conditions, and soil composition. This richdataset serves as input for a CNN model trained onhistorical farm data and weather patterns. ResNet50, withits deep architecture and transfer learning capabilities, isemployed to analyze the collected data and identify keyfeatures indicative of optimal fertilization and pest controlstrategies. The model learns to recognize patterns in plantgrowth, nutrient deficiencies, and pest infestations, enablingit to predict the most appropriate treatments for eachspecific situation. By providing farmers with data-drivenrecommendations, this system aims to enhance crop yieldswhile minimizing chemical usage and reducingenvironmental impact. The solution also incorporatesmachine learning algorithms to continuously improve itspredictions over time, adapting