Automated system and method for optical cloud shadow detection over water

A computer-based system for cloud shadow detection from an image received from a remote sensor comprising:

• index calculator automatically calculating an Integrated Value (IV) index for each pixel in the image;
• ASB creator automatically creating a plurality of electronic spatial ASB, each of the ASBs being selected centered on the pixels to be classified as shadowed or sunlit;
• mean index calculator automatically computing a mean of the IV indices within each of the ASBs;
• cloud remover automatically removing each of the pixels classified as cloud;
• index normalizer automatically normalizing the IV index for each of the selected pixels by the mean of the IV indices within the ASB centered on each of the selected pixels;
• and
• pixel classifier automatically classifying each of the selected pixels as a shadow pixel or a sunlit pixel based on a comparison of the normalized IV index for the selected pixel with a pre-selected threshold.

A non-transitory computer-readable medium storing a computer program for cloud shadow detection from an image received from a remote sensor, said computer program comprising the steps of:

• automatically calculating an Integrated Value (IV) index for each pixel in the image;
• automatically creating a plurality of electronic spatial Adaptive Sliding Boxes (ASB), each of the ASBs being centered on a selected one of the pixels to be classified as shadowed or sunlit;
• automatically computing a mean of the IV indices within each of the ASBs;
• automatically removing each of the pixels classified as cloud;
• automatically normalizing the IV index for each of the selected pixels by the mean of the IV indices within the ASB centered on each of the selected pixels;
• and
• automatically classifying each of the selected pixels as a shadow pixel or a sunlit pixel based on a comparison of the normalized IV index for the selected pixel with a pre-selected threshold.

A computer-based method for cloud shadow detection from an image received from a remote sensor comprising the steps of:

• automatically calculating an Integrated Value (IV) index for each pixel in the image;
• automatically creating a plurality of electronic spatial Adaptive Sliding Boxes (ASB), each of the ASBs being selected centered on a pixel;
• automatically computing a normalizing factor of the IV indices within each of the ASBs;
• automatically normalizing the IV index for each of the selected pixels by the normalizing factor of the IV indices within the ASB centered on each of the selected pixels;
• and
• automatically classifying each of the selected pixels based on a comparison of the normalized IV index for the selected pixel with a pre-selected threshold.