Forestry Hanover Seminar Series presents: Dengsheng Lu, Michigan State University

Date: February 17, 2015
Location: Forestry Hanover Seminar Series presents: Dengsheng Lu, Michigan State University 225 Natural Resources

Time: 4 p.m. (Refreshments at 3:50 p.m.)

Forestry Hanover Seminar Series presents:

Title: The roles of textural images in improving land-cover classification in the Brazilian Amazon

Presented by:

Dengsheng Lu
Center for Global Change and Earth Observations
Michigan State University

 

Abstract: The importance of textures in improving land-cover classification has long been recognized, but how different sensor data with various spatial resolutions affect the selection of textural images is poorly understood. This research examines Landsat TM, ALOS PALSAR L-band, SPOT, and QuickBird with pixel sizes of 30 m, 12.5 m, 10 m, and 0.6 m, respectively, for land-cover classification in the Brazilian Amazon. Grey-level co-occurrence matrix based approaches are used to extract textural images from the above-mentioned sensor data. The maximum likelihood classifier is used to conduct the land-cover classification. This research shows the importance of textural images in improving land-cover classification, and the importance increased as the pixel size improved. Overall, textural images have less capability in distinguishing land-cover types than spectral signatures, but incorporation of textures into radiometric data is valuable. The classification accuracy can be improved by 5.2%–13.4% as the pixel size changes from 30 m to 0.6 m.