舰船尾流红外图像边界检测方法

Border detecting method for IR image of ship wake

由于舰船尾流与周围未扰动的海水相比温差较小且无明显轮廓,导致舰船尾流红外图像成像模糊、对比度低,难以提取其边缘信息。文中根据舰船尾流及其红外图像的特点,提出了一种利用动态纹理和数学形态学检测尾流边界的方法。首先,利用动态纹理分割方法对舰船尾流红外图像序列进行了预分割。然后,采用Canny算子对尾流区进行边缘提取。对提取的边缘利用形态学的方法进行膨胀、标记和选择,提取了真正的尾流边界。实验结果表明:该方法对舰船尾流红外图像序列进行处理,能够有效的提取尾流的边界,取得了较好的处理效果。

Compared with the nearby undisturbed seawater the temperature difference in the ship wake is tiny and the ship wake presents obscure configuration.These induce the low contrast of the infrared ship wake image and the borders of the wake images are difficult to be extracted.According to the characters of ship wake and infrared wake image,a border detecting method by dynamic texture and mathematical morphology was proposed.Firstly,the infrared wake image sequence was pre-segmented by dynamic texture method,and then the edges in the wake area were extracted by Canny operator.The extracted edges were dilated,labeled and selected by the mathematical morphology and the true wake borders were extracted.The experimental results show that processing the wake images by the methods mentioned above can effectively extract the borders of the ship wake and get better results.