基于侧扫声纳的人工鱼礁自动识别方法研究

Research on Automatic Identification Method of Artificial Reef Based on Side Scan Sonar

人工鱼礁的投放是海洋牧场建设的重要措施,为实现大范围人工鱼礁信息的自动识别提取,提出了一套基于侧扫声纳影像的人工鱼礁自动识别方法。首先使用侧扫声纳后处理软件对采集的数据进行预处理,再对预处理后的图像进行均值平滑滤波整体去噪、自适应平滑滤波局部去噪、极差滤波锐化图像、提取目标边缘判定目标、二值化,最后进行矩阵运算提取礁体及其声影区,最终获得鱼礁的位置信息。实验结果表明,本方法应用于两种侧扫声纳图像人工鱼礁识别,正确度达94%以上,完整度达85%以上,且具有良好的通用性,能够为海洋牧场建设中人工鱼礁投放质量评估提供科学的数据支撑。

The deployment of artificial reef is an important measure for the construction of marine pastures.In order to realize a wide range of automatic identification and extraction of artificial reef information based on side scan sonar images,this paper proposes a set of side scan sonar image processing methods.Firstly,this method uses the side scan sonar post-processing software to pre-process the acquired data.Then,a series of processing is performed on the preprocessed image:mean smoothing filtering overall denoising,adaptive smoothing filtering local denoising,range filtering sharpening image,extracting target edge determination target,binarization,etc.Finally,matrix operations are performed to extract the reef and its sound shadow area.After the above steps,the location information of the reef is finally obtained.The experimental results show that this method can identify artificial reefs with more than 94%accuracy and more than 85%integrity based on two side scan sonar images,and with good versatility,which provides scientific data support for the quality assessment of artificial reefs in marine pasture construction.