|
Abstract:
|
The use of remote sensing imagery in precision agriculture has increased dramatically with greater availability of satellites equipped with high-resolution optics (e.g., 5 meters or better). Applications include: predicting end-of-season crop yield maps given in-season remote sensing imagery, in-season detection of growth anomalies, and the use of crop process models that are either driven by inputs derived from remote sensing imagery or use the imagery for data assimilation. However, the use of satellite imagery comes with its own problem in precision agriculture. Detecting cloud/haze contaminated regions in an image is challenging, but vital, and hence calls for effective and robust methods to detect clouds/haze. In addition, high-resolution imagery typically has sparse temporal coverage, which for example is not ideal for crop health monitoring. This problem can be mitigated to some degree by fusing/blending imagery from multiple satellites that have different spatial resolution, temporal coverage, and sensor characteristics. We will given an overview of research efforts at The Climate Corporation that tackle some of these problems.
|
