Since multi-source image classifications have the ability to exceed single source processes, such as traditional unsupervised classification methods, this paper will present the integration of four types of data: Lidar, elevation, multispectral and thermal. Using multi-source data and maximum likelihood classification methodology, as well as all possible permutations of data types, this paper will discuss ways to increase accuracy assessments of forested areas in east Texas and find the best combination of data sources.
Tribby, Hillary; Kroll, James; Unger, Daniel; Hung, I-Kuai; and Williams, Hans Michael, "Multi-Source Image Classification" (2005). Faculty Presentations. Paper 19.