レーダ画像処理による船舶映像の抽出に関する研究

摘要

今回は、レーダ各映像の大きさ、エコーレベル、連rn続性という特徴に着目し、比較的簡単な画像処理を行rnうことで、陸地等の大きな映像を抽出除外し、海面反rn射が多い場所でもその中から船舶と思われる映像を抽rn出することができた。しかし、今回自動抽出した映像rnの中には、まだ船舶以外の映像も含まれている。今後、rn船舶映像のみを抽出するためには、今回抽出した映像rnを追尾し、その動向を観察する必要があると考えられrnる。また、ECDIS等、他の情報と組み合わせることも、rn今後必要となると考えられる。%Ship's navigation officers are required to recognize many ships' movement and to avoid collision risk in sea area with high traffic density. However, radar information is based on reflection of targets. To obtain reliable information of other ships' movement, they have to select target ship's Radar echo from a lot of Radar echoes. If the radar could be discriminate between ships and other images and the automatic acquire function of Automatic Radar Plotting Aid system (ARPA) could be used effectively, ship's officers' workloads would be reduced.rnIn this study, radar echo characters such as size, echo level, and continuous lap were focused and automatic ships image detection algorithm was developed. Continuous lap meant the lap of same position images with superposition of continuous two Radar Images. As a result, the Image processing was enabled to narrow down from 922 to 46 images including 21 ships images.rnThis detective method shows that narrowing down images is available under easy Image processing with short intervals.