In order to join virtual sessions, you must be registered and logged-in(Were you registered for the in-person meeting in Denver? if yes, just log in.) 
Note: All session times are in Mountain Daylight Time.

Incorporating machine learning in urban land use classification based on remote sensing data

Authors: Le Wang, Advisor, Weishan Bai*,
Topics: Land Use, Remote Sensing
Keywords: urban, land use, classification, machine learning
Session Type: Guided Poster
Presentation File: No File Uploaded

Land use of urban area plays a significant role in urban planning and environmental management processes. However, there are several challenges in classification of urban land use by using remote sensing data, such as massive data processing, feature extraction and deriving high-level semantic labels. In order to deal with those challenges, this study tries to incorporate machine learning in classifying detailed urban land use classes based on remote sensing data. This study uses parcels of urban as research units and aims to derive more than seven attributes of land use parcels from remote sensing data to classifying more than four land use classes, civic, industrial, namely office, transportation and educational or medical area. As an experiment, this study uses data of New York city including Lidar data, aerial photographs, land parcel boundary data, and land-use layer data and so on. This study combines all of this data to labialize numerous parcels’ attributes, such as maximum building area, maximum building height, as training data to train models to make classification automatically. As a result, the accuracy expected is 70% and the kappa coefficient expected is 0.55. In additional, the most influential feature of parcels on classification can be figure out after analyzing of result.

Abstract Information

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

To access contact information login