Rice Growth Monitoring Using Small UAV and SfM-MVS Technique

Abstract

 Radio-controlled electric-powered multicopters have become extremely useful as low-cost unmanned aerial vehicle (UAV) with a camera or sensor for proximity remote sensing. Crop growth management is an important use of remote sensing. Particularly for rice, a key crop in Japan, remote sensing is useful to support management for increased yields, optimum harvest time forecasting, and improved eating quality. Observation and prediction have been conducted for wide areas using aircraft and satellite remote sensing. We used a small UAV for growth monitoring of rice: details of growth were monitored using high temporal-spatial resolution images. Then we produced rice growth estimation models. Additionally, structure from motion and multi-view stereo (SfM-MVS) software was used to create ortho mosaic image and crop surface models (CSMs). We used it to monitor rice growth.Observation results revealed details of rice growth differences, which derive from cultivation conditions. Rice growth estimation models yielded good results. We infer that rice production management of techniques such as fertilization can be done using methods used for this study.