摘要: 运用傅里叶分析方法,利用水平方向、垂直方向、45°方向和135°方向直线的频域模板对侧扫声纳图像在4个方向上进行投影分解,根据子图像在4个方向上投影的能量和,设定门限判断有无目标,并粗略估计目标的位置。实验结果证明,该方法对于纹理较强的侧扫声纳图像中感兴趣区域(目标)的检测有较好的效果。
关键词:
傅里叶分析,
频域模板,
投影,
感兴趣区域,
侧扫声纳图像
Abstract: This paper presents a Fourier analysis method for target detection in side-scan image. Using horizon line, vertical line, slope 45° line and slope 135° line spectral mask, the projection of side-scan sonar image in these orientation is obtained. Based on the sum of energy of four orientation projections, a threshold is set to judge target existing, and the position of target is also given. Results show that the method has a perfect performance for regions of interest (target) detection of side-scan sonar image with strong texture.
Key words:
Fourier analysis,
spectral mask,
projection,
regions of interest,
side-scan sonar image
中图分类号:
罗明愿, 卞红雨, 周志娟. 基于图像频域方向模板的目标检测方法[J]. 计算机工程, 2011, 37(12): 215-217.
LUO Meng-Yuan, BIAN Gong-Yu, ZHOU Zhi-Juan. Target Detection Method Based on Image Spectral Direction Mask[J]. Computer Engineering, 2011, 37(12): 215-217.