Geographic Window Sizes Applied to Remote Sensing Sea Surface Temperature Front Detection

Abstract

The effects of using a geographic window size with an existing edge-detection technique for the detection of thermal fronts in sea surface temperature (SST) imagery are investigated. The size of a geographic window is not constant but instead is determined by the correlation of the data surrounding the window's central point. Using this approach instead of a fixed window size, the investigation windows are optimized for the whole image, providing more reliable detection of edges within the windows. The new algorithm was run on several SST images from southern Lake Michigan and compared to runs of the original algorithm and a modification of the original algorithm optimized for this region. The results show that the geographic window improves edge detection most in the nearshore regions and to a lesser extent in the offshore regions.