CSEWIS: Cloud and Snow Estimates from Webcam Image Streams
AuthorMunoz, Hannah J.
Computer Science and Engineering
AltmetricsView Usage Statistics
Environmental monitoring is often accompanied by inexpensive camera technology to obtain automatic time series of of site conditions. However, the resulting pictures will often go unused for advanced analysis because of the lack of fast and reliable ways to quantify data from images. This thesis presents CSEWIS, an image analysis software that aims to help fill the gap of missing analysis software by delivering accurate, repeatable coverage estimates of snow or clouds for use in environmental research. CSEWIS uses an image series from remote sensor networks and allows users to select regions of interest within the image and perform snow or cloud coverage analysis. Analysis algorithms can be done on the CPU, GPU, or on multiple GPUs for faster execution times. Results are delivered as a CSV file with a related XML metadata document. An optional AVI video file of the coverage can also be exported. By producing quantifiable results, CSEWIS aims to increase the usability of webcams commonly used in sensor networks. This thesis details the background behind image processing, remote sensing, and webcams, overviews the system architecture, and how to use the resulting product. A comparison to related works and future work on the project are discussed.