A Detection Method of Ships through a Navigational Image Sequence taken by Multi-Sensor Camera System
作者单位：Tokyo University of Marine Science and Technology 2-1-6 EtchujimaKoto-KuTokyo 135-8533Japan Tokyo University of Marine Science and Technology 2-1-6 EtchujimaKoto-KuTokyo 135-8533Japan Tokyo University of Marine Science and Technology 2-1-6 EtchujimaKoto-KuTokyo 135-8533Japan Tokyo University of Marine Science and Technology 2-1-6 EtchujimaKoto-KuTokyo 135-8533Japan Tokyo University of Marine Science and Technology 2-1-6 EtchujimaKoto-KuTokyo 135-8533Japan
会议名称：《Asia Navigation Conference 2005》
会议日期：2005年
学科分类：08[工学] 082402[工学-轮机工程] 0824[工学-船舶与海洋工程]
摘 要：正According to the statistic of Japan Coast Guard, in 2003 the numbers of marine accidents in Japan’s waters were 2500 or more. The marine accidents mostly occurred within 3 nautical miles from the coastal line or in the port were collision or stranding. Meanwhile 75% of the marine accidents were caused by human error, due to insufficient lookout. To deal with this problem, we suggest the application of image-processing techniques to automate the lookout system. The purpose of this system is to prevent oversight of lookout, therefore the marine accidents can be minimized and the safety and efficiency of maritime vessel traffic can be assured. We proposed a detection method of moving ships through a navigational image sequence taken by cameras installed at ship’s bridge. Since the image is influenced by roll and pitch, the general image processing technique, such as Background Difference method cannot be used. Moreover, in the images classification sea waves appear as noise. So we proposed a novel technique to eliminate the influence of roll, pitch and sea waves in the images. Since the application of the video camera causing insufficient exposure to light during the night, or the detection of ships was difficult under rain condition, this system used the infrared camera together with the video camera. We verified the advantage of using the multi-sensor camera system composed of the video camera with CCD （Charge Coupled Device） sensor and the infrared camera with bolometer sensor under rain condition in the twilight at Tokyo Bay.