Abstract
Digital image processing technique and fuzzy logic approach are used to identify forest areas infested with Southern Pine Beetle, SPB, using normal color airborne imageries in this research. This research will be used as a front end of a larger framework, Southern Pine Beetle Visual Information System (SPBVIS), that integrate mathematical model, weather station data and airborne images. SPBVIS can be used by forest managers to predict the direction and the rate of infestation of SPB by using spatial information obtained from airborne images to drive the Southern Pine Beetle simulation model. Color airborne imageries obtained from S-VHS video signal are a normalizing color transform is applied to the input image's color space. Dynamic fuzzy logic membership functions are then used to threshold the resulted color space. All the pixels then classified as either belonging to or not belonging to the infestation spot. Event Probability Correlation Analysis (EPCA) and the percentage of correctly classified pixels are used as indicators to measure the performance of the segmentation algorithm. Images clustered by domain experts are used to compare to the images generated by the segmentation algorithms.
Ng, Kit-Tong (1994). Fuzzy logic approach to supervised segmentation of forest regions infested by Southern Pine Beetle using color airborne images. Master's thesis, Texas A&M University. Available electronically from
https : / /hdl .handle .net /1969 .1 /ETD -TAMU -1994 -THESIS -N575.