Agricultural and Land Management Using AI: A Case Study of Rice Plot Identification in Senegal

Agricultural yield improvement is important to handle food insecurity mostly for developing countries. Indeed accurate knowledge of the distribution of crops in the landscape is crucial for better management and monitoring of the agricultural sector. In recent years the combination of artificial intelligence (AI) and remote sensing data has been widely used for crop type mapping. Given the essential place of rice in the Senegalese diet, increasing its production can positively impact food security. Thus, having an estimate of its harvests can be useful to stakeholders for better management of the rice sector in Senegal. In this work, we aim to build an AI system with remote sensing data for rice crop mapping in Senegal. We exploit two AI models with Sentinel-2 images for rice mapping. The first model is based on Support Vector Machine (SVM) and a second model based on deep learning using the Deeplab V3+ model. Both models shows promising results even if they still very low. The results reveals that the deep learning model provides better performance at identifying rice crop than the SVM model which has a lower accuracy.