Comparison of SIED and Canny Algorithm in Caspian Sea Surface Temperature Front Detection Using Satellite Image

Sea surface temperature fronts, narrow strips on the ocean's surface with significant temperature changes, play a crucial role in marine ecosystems and climate regulation. This study compares the Single Images Edge Detection (SIED) and Canny algorithms in detecting sea surface temperature fronts in the Caspian Sea using MODIS satellite images from 2015 to 2019. The SIED algorithm, a population-based method, identified fronts by statistically analyzing temperature histograms within a 32×32-pixel window. In contrast, the Canny algorithm, a gradient-based method, detected fronts by calculating temperature gradients at each pixel. Both algorithms revealed seasonal and spatial variations in temperature fronts, with the highest presence of fronts detected during the winter months. The SIED algorithm found the lowest presence of stable fronts in the northern Caspian in September and April and the southern Caspian in November, March, and April. The Canny algorithm showed the lowest presence in June, March, and August. SIED detected the highest presence of stable fronts in November and December in the northern Caspian Sea and in January in the southern Caspian Sea. The Canny algorithm identified the highest presence during the first three months of the winter monsoon. Both algorithms consistently detected fronts along the eastern coasts of the Middle and South Caspian, with significant fronts near the Garabogazköl Basin and Turkmenbashi Gulf. Despite differences in detection, both methods revealed similar general patterns of temperature fronts