Abstract

Objective:
    To analyze soy production in Kansas using GIS.
Summary:
    In order to create an image emphasizing soy a comparison must be made between the unique spectral signature of soy and the spectral bands being used by the satellite scanner.  Soy's spectral signature shows a 50% peak reflectance in the near infrared band at 0.8µm and tapering off slightly into the mid-infrared band at 1.3µm.  Upon comparison it is noticeable that soy's peak reflectance matches up with the last 75% of Thematic Mapper Band 4, which gathers light in the near-infrared spectrum from 0.76µm to 0.9µm.  There is a gap of 0.65µm between band 4 and 5 where soy's reflectance continues to peak, but is not recorded by the Scanner.  Band 5 begins at 1.55µm.  This, however, will not negatively affect the study.   Another method of matching band combinations to specific spectral signatures is to use the Spectral Characteristics Viewer on the USGS website (not yet applicable for soy). 
    For the purpose of analyzing soybeans, a false color composite image will be created using TM band 4 displayed as red, TM band 5 (moisture) will be displayed as green and TM band 3 (visible red) will be displayed as blue for crop classification.   Knowing that the phenological cycle of soybeans is May through October, I expect to see very little in the May scenes.  I expect to see abundant vegetation with bright colors denoting soybeans in July and representation in darker colors for the September dataset when the photosynthetic activity begins to slow down.  Because soybean's spectral signature continues peaking right off the chart in the near-infrared spectrum I would anticipate it to display with the highest levels of visible red.  The higher levels of red being orange and yellow.
Conclusion:
    Because soy is planted and harvested a month later than corn, it will be possible to define the crops remotely by studying the maps for planting and harvesting activities.  A very observable activity for the satellite scanners.
 
 

 

ES 775 Advanced Image Processing, Student Project by Matthew Randall
 Emporia State University
, James Aber - Instructor

 
Comments