Subjective and Objective Evaluation of Data Quality Visualization Methods on Navigational Charts

TitleSubjective and Objective Evaluation of Data Quality Visualization Methods on Navigational Charts
Publication TypeConference Proceedings
AuthorsKastrisios, C, Ware, C
Conference NameAutoCarto 2022
Conference DatesNovemebr 2-4
Conference LocationRedlands, CA

Nautical charts are compiled with bathymetric data that differ in the year and method collected over areas varying from stable to mobile seafloor. Data quality on Electronic Navigational Charts (ENCs) is mainly encoded as an overlay of polygonal regions, each assigned a Category Zone of Confidence (CATZOC) based on the positional and attribute accuracy of depth measurements and the completeness of the collected data. With CATZOC seafarers may estimate where it is safe to navigate, including areas of minimal underwater clearance, and where improved awareness is required for the presence of uncharted, or poorly charted, hazards. Failing to account for the varying data quality may result in maritime accidents, environmental disasters, and loss of life (e.g., (BSU, 2020; DSB, 2017; RMIMA, 2020)) Nautical cartography is undergoing major changes with the development of new data standards and products based on the S-100 framework (IHO, 2018; Kastrisios et al., 2022). This includes the transition from CATZOC to the new composite data quality indicator for ENCs, the Quality of Bathymetric Data (QoBD). The CATZOC alphanumeric codes (A1, A2, B, C, D) are replaced by a numerical scheme (1 for best quality and 5 for worst) with one more category O (“Oceanic”) for areas of depths greater than 200m (“U” for unassessed remains). Furthermore, the relevant committees of the International Hydrographic Organization (IHO) have decided on replacing the current star symbology that has been proven problematic (see Kastrisios et al., 2020). One potential solution for the data quality sectors’ visualization is with the use of seethrough textures consisting of countable elements. In this work we discuss the proposed coding schemes (Lines and Dot-Cluster) and the results of an online user survey and an in-lab experiment for their evaluation compared to three alternatives (Opaque-Colors, Color-Transparency, and Color-Textures). Furthermore, we discuss how the results compare to previous works on uncertainty visualization and how the concept of countable textures could be extended for other uses.

URL for Proceeding