A Simple Thermal-Front Detection for Cloudy Sea Surface Temperature Data

The detection of thermal-fronts using data derived from satellites imagery data tends to be unsuccessful when the data being analyzed are largely covered by cloud. Segmentation or clustering-based detection method requires clear data and large size of dimension to achieve success on segmentation or clustering process. Otherwise, it needs a complex pre-process stage to process the cloudy data before the detection process. We proposed a simple thermal-fronts detection to process the Aqua MODIS Level 3 data that have a lot of clouds cover. By using the mathematical approach, we developed a model based on the pixels detection by analyze the values of the pixels and its relationships to neighbor pixels. We applied the segmentation approach using two window masks, i.e., searching window and neighboring window. Each pixel in the searching window will be evaluated based on the difference value (DV) between the center pixel and its' neighbor pixel, meanwhile each neighbor pixels will be created a neighboring window and evaluated using the edge thresholding value (ETV). The results from DV and ETV are used to determine whether or not the pixels in the searching window can be set as the thermal-fronts pixels. The experimental results show that the proposed method has a good performance to detect thermal-fronts using cloudy data.