Using Deep Learning for Evaluating and Managing Crop Growth Condition based on Multi-Source Images

Image-based assessment of crop conditions is one of the most important and basic components of precision agriculture. It is widely used to evaluate and predict crop growth and yield. In addition, the evaluation and prediction results can further be used for many other aspects of crop production, such as decision-making for water and fertilizer management, evaluation of sowing quality, and assessing seed quality. Therefore, it is necessary to build a comprehensive assessment solution for crop growth conditions based on big data. Big data of image sensing plays an important direction in precision agriculture. However, to the best of our knowledge, there is no such a precision agriculture scheme that can systematically make an integrated evaluation and decision based on the big data of image sensing. In this research, we will leverage various Deep Learning (DL) solutions to build such a system that can (1) select effective information from multi-source sensing data; (2) comprehensively evaluate crop growth conditions; (3) support the decisions of crop management.