Soybean farmer mobile application

Abstract The demand for soybean tends to increase worldwide with population growth. Despite farmers adopting strategies such as crop rotation, soil preparation, and the application of chemical products, the most critical factor to increase productivity is choosing cultivars adapted to the region so that they are resistant to deceases. In addition, seed quality also directly influences crop productivity. To help a farmer select suitable cultivars and analyze the quality of soybean seeds, we present an application that uses the user’s location to identify suitable cultivars based on the edaphoclimatic characteristics of a region. The details of these cultivars are shown in text, images, and video, making it possible to compare the selected varieties. Our app also contains a trained convolution neural network capable of classifying the quality of soybean seeds based on an image captured by the user or stored on a smartphone. The convolutional neural network architecture allowed an excellent performance, with an accuracy of 94.06% in the classification of soybean seeds. All app functionality runs comfortably on mobile devices. Compared to others that have the same purpose, our application has a more significant number of features.