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.

Planet Satellite Image Segmentation and Classification

Authors: David Mills*, Texas State University
Topics: Environmental Science, Geographic Information Science and Systems, Remote Sensing
Keywords: machine learning, SLIC, quickshift, Felzenszwalb, watershed, remote sensing, satellite imagery, planet, climate, air quality, EPA
Session Type: Paper
Presentation File: No File Uploaded


This research uses satellite imagery courtesy of Planet Labs to search for correlations between local construction projects and EPA air quality readings. First, mostly cloud-free satellite imagery was gathered using the location of EPA air quality monitors as a center point. Then, the imagery was segmented using Felzenszwalb, Quickshift, Watershed, and SLIC algorithms. A percentage of the data was used to train the algorithms to appropriately classify pertinent land use changes such as the clearing of land for construction and development. After the segmentation and classification, dates which indicated construction or other relevant land use changes occurring were tested against fluctuations in average PM 2.5 and PM 10 readings for the same dates at EPA Air Quality Monitoring sites. The influence that local land use changes have on air quality monitor readings have implications for the general usefulness of EPA air quality monitors. As cost-efficient, high quality, crowd-sourced internet-of-things air quality monitoring technologies continue to improve and be distributed at an increasing pace- the is necessary to question the capability and benefit of traditional and expensive EPA air quality monitors.

To access contact information login