| 摘要 |
<p>The present invention estimates harvest with high accuracy using aerial images and time series weather data during specified crop growth stages while keeping the number of inspected fields to a minimum. From previously collected images and field attribute information for the inspected region and aerial images taken of the fields that are the target of the inspection, a parameter group to serve as determination criteria for selecting measurement fields that should be inspected is determined. The measurement fields are selected so that the parameter group has variance whenever possible. To reduce the burden of the inspection as much as possible, the measurement field candidates are selected so as to be locationally concentrated as much as possible. By analyzing the weather data time series patterns for each growth stage, a parameter group that correlates to the growth conditions is calculated and harvest is estimated using image feature values, field attribute information and the parameter group as explanatory variables.</p> |
