When Chemistry is Too Colourful: Gamut Clipping in 8-bit sRGB Risks Misinterpretation of Camera-Based Chemical Analysis [machine readable data]

Digital cameras are increasingly utilised to capture visual changes in chemical processes. Monitoring colour with computer vision tools serves as a valuable proxy for monitoring bulk chemical changes. Most consumer-grade cameras digitise the colours captured using the 8-bit sRGB colour space. Despite its ubiquity, it restricts the range of colour information that can be stored. With regards to chemical analysis and process monitoring, the limitations of using the 8-bit sRGB gamut have not been addressed. When real-world vivid colours one attempts to record lie outside the bounds of 8-bit sRGB, the digitisation of those colours can result in distortion, clipping, or loss of chemically relevant colour data. Ultimately, these sRGB gamut limitations risk the chemist misinterpreting the data they collect from a camera. In this paper, we examined the visible spectrum of a series of common dyes, and determined their colours spectroscopically, without the limitation of 8-bit sRGB encoding. We investigated how 8-bit sRGB encoding affects the interpretation of time-series data from theoretical colour changes in five dyes. Highly saturated chemical samples exceeded the 8-bit sRGB colour gamut, causing colour distortions and structural breaks in reaction-monitoring time series data, risking misinterpretation as kinetic phenomena of genuine chemical origin. Our findings underscore the importance of paying close attention to colour representation in digital chemistry. We offer practical guidance for researchers using and interpreting colour data for use in computer vision method development.

数码相机正日益广泛地被应用于捕捉化学过程中的视觉变化。借助计算机视觉工具监测颜色，可作为监测整体化学变化的有效替代指标。大多数消费级数码相机采用8位sRGB（8-bit sRGB colour space）色彩空间对采集到的颜色进行数字化处理。尽管该色彩空间应用极为广泛，但它限制了可存储的颜色信息范围。针对化学分析与过程监测场景，8位sRGB色域的局限性尚未得到充分解决。当研究者尝试记录的真实高饱和度颜色超出8位sRGB色域范围时，这些颜色的数字化过程可能会出现色彩失真、色彩截断，或是丢失与化学特性相关的颜色数据。最终，sRGB色域的这些局限性可能导致化学家对相机采集的数据产生误判。在本研究中，我们对一系列常见染料的可见光谱进行了分析，并通过光谱学方法测定了它们的颜色，未受8位sRGB编码的限制。我们还探究了8位sRGB编码会如何影响五种染料理论颜色变化的时序数据解读。高饱和度的化学样品颜色会超出8位sRGB色域，进而导致反应监测时序数据出现色彩失真与结构断点，可能被误判为源自真实化学过程的动力学现象。本研究结果凸显了在数字化学领域重视颜色表征的重要性。我们还为从事计算机视觉方法开发、需使用并解读颜色数据的研究人员提供了实用指导。