关键词: 警示服饰, 警示色, 可视性, 标准, 图像分析

Abstract: The effective evaluation of color usage in warning garments can enhance the color design effectiveness, playing an important role in preventing accidents and protecting human life safety. To clarify the evaluation methods for warning garment colors, we focused on reviewing relevant research on the evaluation of warning garment color effects, and made a summary from the two aspects of warning color evaluation methods and related standards, providing reference and development direction for the color design and warning effect evaluation. The evaluation methods for warning garment color usage mainly include subjective investigation, instrument measurement, and image analysis. Subjective investigation and instrument measurement mainly evaluate from individual psychological or physiological perception, and the individual biases of participants have different degrees of impact on the evaluation results. Furthermore, the subjective investigation method and the targeted arrangement of the instrument measurement method inhibit the feasibility of future reuse. In contrast, the image analyzing method describes the differences in machine vision between clothing and scenes with focus on simulating human common perception through machine vision, and also provides feasible solutions for batch evaluation and analysis, which can achieve efficient and convenient commercial evaluation to a certain extent. However, the disadvantage lies in the fact that it ignores the impact of cognitive experience on the risk perception of participants due to different social status and living environment, i.e., the true feedback from participants. Overall, the existing methods are difficult to balance the convenience and the consistency of subjective and objective evaluation results. Therefore, it is necessary to optimize and improve the theory, research methods, and evaluation application from various aspects based on the existing methods. It is feasible to improve the robustness and accuracy of image analysis in changing environments, combine the advantages of convenient and immediate image analysis with the advantages of subjective investigation and instrument measurement closely related to participants to achieve objective evaluation consistency, and utilize deep learning technology and participant feedback datasets to make computers learn common feedback from people. Furthermore, this paper analyzes in-depth the factors influencing the development trend of warning color evaluation methods, pointing out that there is still room for further optimization in terms of universality, consistency of subjective and objective evaluation, and the fit of clothing products. Achieving rapid evaluation technology in complex backgrounds, improving the consistency of subjective and objective perception in evaluation methods, and the compatibility of style modeling and color are feasible development directions for warning color evaluation of clothing in the future.