One of the first steps in evaluating fisheries acoustic data is to explore an echogram of the data. Echograms are a visual representation of reflected sound with range or distance from the echosounder on the $y$-axis, time on the $x$-axis, and the intensity of the reflected sound displayed through a mapped color scale. Many of the traditional color scales used in fisheries acoustic echograms have not changed in the last few decades and have been previously discussed in the literature. However, little research has been conducted on how the number of colors used, and color thresholds and limits influence our perception of patterns in acoustic data. Understanding how color maps influence out perception bias is particularly important when the analyst is unfamiliar with the dataset, the dataset has areas of dense, overlapping target aggregations, or the data are noisy. This study collected feedback on four different datasets from fisheries acoustics experts to understand how colormap settings influence perception of echograms. The datasets included a simulated, noisy echogram, an echogram with both volume and point targets, an echogram with layers of different intensities, and a high-frequency echogram affected by background noise. Colormap settings based on data quantiles, both for thresholds and color limits, were chosen by most of the experts as the echograms that best represented the data for the given task. This study recommends histogram-equalized colormaps for initial data exploration. General recommendations are provided based on results.

评估渔业声学数据的第一步之一就是探索数据的回波图。声波图是反射声音的视觉表示，具有与回声测深仪之间的距离或距离。$ÿ$轴，时间在 $X$轴，以及通过映射的色标显示的反射声音的强度。在过去的几十年中，许多用于渔业回声图的传统色标没有变化，并且先前已经在文献中进行了讨论。但是，关于使用的颜色数量以及颜色阈值和限制如何影响我们对声学数据中的模式感知的研究很少。当分析人员不熟悉数据集，数据集具有密集的重叠目标聚集区域或数据嘈杂时，了解色彩图如何影响感知偏差尤为重要。这项研究收集了来自渔业声学专家的四个不同数据集的反馈，以了解色图设置如何影响回波图的感知。数据集包括一个模拟的，有声的回波图，包含体积和点目标的回波图，具有不同强度层的回波图以及受背景噪声影响的高频回波图。大多数专家基于阈值和颜色限制的基于数据分位数的颜色图设置被选作最能代表给定任务数据的回波图。这项研究建议使用直方图均衡的颜色图进行初始数据探索。根据结果​​提供一般建议。这项研究建议使用直方图均衡的颜色图进行初始数据探索。根据结果​​提供一般建议。这项研究建议使用直方图均衡的颜色图进行初始数据探索。根据结果​​提供一般建议。