基于数学形态学的水声图像处理

Underwater acoustic image processing based on mathematical morphology

对成像声纳采集到的水声图像进行有效且快速的处理与正确识别,是实现水下智能机器人的“声视觉”的关键问题。针对水声图像的特点,传统方法难以保证对“保边去噪”的要求和保持系统算法整合性。为此提出了一种基于数学形态学的水声图像处理方法,该方法采用广义形态开闭变换等形态学方法进行图像处理。试验结果表明,该算法抗噪能力强,效果明显,且计算量小、硬件实现简单,更好地保证了“声视觉”系统的准确性和实时性。

It is important for the Acoustic Vision System of AUV to process and recognize the underwater acoustic images obtained by imaging sonar effectively and fleetly. Considering the features of acoustic images, traditional methods can hardly guarantee the effect of both edge-reserved and noise-smoothing and conformity of systemic algorithms. In this paper, a method of underwater acoustic image processing based on mathematical morphology with broad-sensing open-close or so is proposed. Experiments show that this ...