Hyperspectral remote sensing image analysis for citrus pest monitoring

The field of remote sensing has undergone tremendous development in the past decades. One of the fastest growing applications of remote sensing is precision agriculture. Perhaps no one area of remote sensing has had as many implications for the environmental and agricultural studies as hyperspectral imaging.; Hyperspectral imaging data provides high spatial and spectral resolution information which is used to estimate tree parameters that can be in turn used to detect pest stress in the trees. This information not only helps detect pest infestation, but it is also useful in extracting information about the type of infestation, and the extent of the same. This makes it possible to continuously monitor the health of trees and maximize profits, by maintaining good quality and high nutritional value. The choice of image analysis methods is critical for hyperspectral image analysis, because the method should be capable of taking proper advantage of the abundance information contained in the image. In this thesis, simultaneous individual tree crown delineation and treetop detection was chosen because the results obtained are suitable for further analysis in extraction of tree parameters for pest-monitoring.