Fine crop mapping by high spatial resolution remote sensing data in complex heterogeneous areas

Authors: Shougeng Hu, supervisor, Peng Zhang*, China University of Geosciences
Topics: China, Land Use, Remote Sensing
Keywords: Wordview-2, Crop mapping, Heterogeneous area, Random forest classifier, Remote sensing, object-based classification
Session Type: Illustrated Paper
Day: 4/11/2018
Start / End Time: 3:20 PM / 5:00 PM
Room: Canal St. Corridor, Sheraton, 3rd Floor
Presentation File: No File Uploaded


Crop classification of complex heterogeneous landscapes is a common requirement to estimate land cover mapping, monitoring, and land use categories accurately.On the outskirts of China's big city, the fragmentation of arable land is serious and the crop planting structure is complex.The objective of this study was to assess the prospects of mapping the common agricultural crops in highly heterogeneous study area in Urban suburbs using high spatial resolution World-view 2 imaging data. Minimum noise fraction transformation was used to pack the coherent information in smaller set of bands and the data was classified with random forest (RF) algorithm and Classification and Regression Trees (CART) algorithm . A total of 5 crop species were mapped in the field. The RF classifier achieved a high overall accuracy (83.02%), whereas the CART classifier produced a lower overall accuracy(81.20%).This study demonstrated that fine crop mapping in complex heterogeneous areas can be made using high-resolution images from a single season, resulting in spatially consistent results. Further,The land cover classification in Wuhan region shown that the Random forest classifier and Classification and Regression Trees classifier compared to no obvious advantage.

Abstract Information

This abstract is already part of a session. View the session here.

To access contact information login