摘要
对船舶水尺视频图像的特点进行分析,设计吃水计算算法。采用纹理谱方法,提取相应的视频帧图像对比度和方向度特征,通过聚类算法进行分类,实现对船舶水尺图像的分割,从而确定出船舶吃水线的吃水准确数值。实验表明该方法速度快,精度高,可以解决港口船舶自动计重的吃水数值准确确定问题。
Ship waterline is at the junction of water and shipboard regions, and precise positioning of ship waterline affects the accuracy of ship load calculation directly. For the characteristics of ship waterline image, features of BWS and DD are extracted based on texture spectrum method,then ship sub-images are classified by K-means clustering algorithm in order to segment ship waterline images. As a result, the location of waterline is determined quickly and accurately. Experimental results show that proposed algorithm has high speed and accuracy during precise positioning of ship waterline.
出处
《皖西学院学报》
2009年第5期1-4,共4页
Journal of West Anhui University
基金
安徽省高校自然科学基金重点项目资助(KJ2009A54)
关键词
纹理谱
船舶水尺图像
特征提取
黑白对比度
方向度
texture spectrum
ship waterline image
feature extraction
black and white contrast
degree of direction