In 2000 the District received a grant from the Washington Department of Ecology to develop a method to identify crops grown during the summer by the use of satellite images. The alternative, to visually identify crops, requires an enormous amount of manpower time to view each of these fields.
The process works by visually identifying the crops in approximately ten percent of the fields at the time a satellite is taking an image, then using that information to create image-signatures of each type of crop. The signatures are then used to identify the remaining crops in the image.
The District has determined that a minimum of two images per summer must be used to create signatures unique enough between crops. For example, the image signature of fields growing wheat and alfalfa are not greatly different during May, but are very different in July when wheat has turned gold and/or has been cut. The combination of two or more images allows different signatures for the majority of major crops in Franklin County.
This process was initially applied in 2001 with an accuracy of approximately 85%. The process was repeated with some modifications in summer 2002 and an accuracy of approximately 87%. The accuracy varies by crop -- potato has a signature relatively unique, making it the most accurately classified at about 95% each year.
The process will again be done during summer 2003. The information is used to support resource management issues such as locations to apply improved irrigation management.
How it Works
Remote sensing is the observation and measurement of objects from a distance, i.e. instruments or recorders are not in direct contact with objects under investigation. Remote sensing depends upon measuring some kind of energy that is emitted, transmitted, or reflected from an object in order to determine certain physical properties of the object. One of the most common types of remote sensing is photography.
These techniques are based on sensing electromagnetic energy emitted or reflected from the Earth's surface and detected at some altitude above the ground. The electromagnetic spectrum is, thus, the starting point for understanding remote sensing. Passive remote sensing is based on detecting available (background) electromagnetic energy from natural sources, such as sunlight. Active remote sensing, in contrast, depends on an artificial "light" source, such as radar, to illuminate the scene.
