A Marine recreational vessel reconnaissance system utilizing IKONOS imagery
Abstract
This dissertation investigates the ability of IKONOS imagery to detect small recreational boats. To accomplish this, automatic target detection software called MRV Recon has been developed which makes use of a weighted Euclidean distance metric. To test the detection accuracy of MRV Recon, a dataset was created by gathering position and attribute data for 53 recreation vessel targets within Cadboro Bay, British Columbia, Canada. IKONOS imagery was collected in May 2003.
The overall detection accuracy was 77 %. The targets were broken down into two categories: A) less than 6 m in length, and B) greater than 6 m long. The detection rate for the category B targets was 100%, while the detection rate for the category A targets was 61%. It is important to note that some category A targets were selected specifically to test the detection limits of MRV Recon. The smallest target detected was 2.2 m long and 1.1 m wide. The analysis also revealed that the ability to detect targets between 2.2 m and 6 m long was diminished if the target was a dark colour. It has been demonstrated that MRV Recon will provide the Canadian Coast Guard with a unique and effective tool for gathering crucial data on recreational vessels.