The routine assessment and monitoring of aquatic habitat characteristics, such as submerged aquatic vegetation (SAV), is a fundamental tool in fisheries management. The relative low cost and availability of recreation‐grade side‐scan sonar (SSS) systems has made capturing high‐resolution (<10 cm) underwater imagery a practical and viable tool for improving these assessments. Using Humminbird SSS and a boat‐mounted transducer, we determined our ability to produce classified maps of SAV within three reservoirs in Texas. Sonar recordings of photic zone habitat were collected and processed into georeferenced mosaic images by using SonarTRX SSS software for use in geographical information systems (GISs). The resulting raster images were interpreted to identify and map SAV, and a classification scheme was developed. Random‐point data was also used to develop a traditional error matrix and an accuracy assessment for each classified map. In the three reservoirs, 485 ha were classified, including 298 ha of SAV. The area estimates for 11 of 12 classes fell within the 95% CIs that were estimated from concurrent random‐point surveys. Overall classification accuracy ranged from 78% to 88% for the three reservoirs. The methods that were developed demonstrate an accurate procedure for calculating SAV coverage and producing a high‐quality map product for distribution to stakeholders.

中文翻译：

---

使用娱乐级侧面扫描声纳绘制水下水生植被的分类图

例行评估和监测水生生境特征，例如淹没的水生植被（SAV），是渔业管理中的基本工具。娱乐级侧面扫描声纳（SSS）系统的相对较低的成本和可用性使捕获高分辨率（<10 cm）的水下图像成为改进这些评估的实用且可行的工具。通过使用Humminbird SSS和船载传感器，我们确定了在德克萨斯州三个水库内制作SAV分类地图的能力。使用SonarTRX SSS软件收集光合带栖息地的声纳记录，并将其处理成地理参考的马赛克图像，以用于地理信息系统（GIS）。解释生成的光栅图像以识别和映射SAV，并开发了分类方案。随机点数据还用于开发传统的误差矩阵和每个分类图的准确性评估。在这三个水库中，分类了485公顷，包括298公顷的SAV。12个类别中11个的区域估计值落在通过并发随机点调查估计的95％置信区间内。三个储层的总体分类精度范围从78％到88％。所开发的方法证明了计算SAV覆盖率和生产高质量地图产品以分发给利益相关者的准确程序。三个储层的总体分类精度范围从78％到88％。所开发的方法证明了计算SAV覆盖率和生产高质量地图产品以分发给利益相关者的准确程序。三个储层的总体分类精度范围从78％到88％。所开发的方法证明了计算SAV覆盖率和生产高质量地图产品以分发给利益相关者的准确程序。