A Novel Framework Based on Deep Learning and Unmanned Aerial Vehicles to Assess the Quality of Rice Fields

In the past few decades, boosting crop yield has been extensively regarded in many agricultural countries, especially Vietnam. Due to food demands and impossibility of crop-field area increasing, precision farming is essential to improve agricultural production and productivity. In this paper, we propose a novel framework based on some advanced techniques including deep learning, unmanned aerial vehicles (UAVs) to assess the quality of Vietnamese rice fields. UAVs are responsible for taking images of the rice fields at low or very low altitudes. Then, these images with high resolution will be processed by the deep neural networks on high performance computing systems. The main task of deep neural networks is to classify the images into many classes corresponding to low and high qualities of the rice fields. To conduct experimental results, the rice fields located in Tay Ninh province are chosen as a case study. The experimental results indicate that this approach is quite appropriate for agricultural Vietnamese practice since its accuracy is approximately 0.72.