GIS-enabled maritime awareness system (GMAS)

Being aware of maritime activities is important for homeland security. We have worked with the US Coast Guard to develop a waterway visual surveillance system on Elba Island, in Savannah, GA for maritime awareness. With the system, we developed an enhanced temporal difference algorithm based on active zone (ETD-AZ) to detect waterway moving objects, including container ships, commercial vessels, small vessels, sailboats, tugboats, and dolphins, etc. In the algorithm, active zone is first defined to model the imaging locations of moving objects that appear in the waterway; the model reduces noises caused by winds, clouds, and waves in the near-view fields. A thresholded difference map is computed by using temporal difference and thresholding methods. The thresholded difference map is further processed by applying morphological operations to remove noises. A final decision rule is derived to decide whether an image contains moving objects or not based on motion pixels in thresholded difference map and in the difference map after morphological operation. The proposed ETD-AZ algorithm has been tested using 21, 844 night and day images taken from Aug. 1-2, 2008, in different weather conditions: sunny, cloudy, windy, stormy with lightening, etc. The algorithm detected 89 of 94 waterway moving objects with 5.3% false negative rate while maintaining a low false positive rate of 4.0% for a total of 21, 478 non-object images. When focusing only on vessels, it detected 78 of 79 moving vessels with a 1.3% false negative rate.