Color-based analysis has been widely used in many resource-limited situations. However, one of the challenges for this type of application is to quantify the color information, especially when the color changes and shade changes occur together. In this work, pH sensing with paper strips, where both the color and the shade vary with changing pH and are easily influenced by ambient illumination, was used as an example to develop a color quantification method based on image analysis. The images were obtained by photographing the pH strips in aqueous solution. The color information was then extracted from the images, and its relationship with pH was explored in both red, green, blue (RGB) and hue, saturation, value (HSV) color spaces. It was found that the hue of the HSV space has a good correlation with the pH, despite changes in both the shade and the color of the sensing strips. The color of the sensing strips was then quantified by transforming the color to hue and a reverse model to predict the pH of water was established. It was found that the R² of the model is 0.952 and the root mean square error (RMSE) is 0.219, better than the visual interpretation precision of pH strips (0.5). Finally, the applicability of the model under different lighting conditions was discussed. Thus, color quantization with both color and shade changes has been achieved with high accuracy. We believe that this could find wide application in various point-of-use devices where both color and shade changes are encountered, such as in outdoor scenarios.

中文翻译：

---

比色法和图像分析中组合的颜色和阴影变化的量化：以水pH值测量为例†

基于颜色的分析已在许多资源有限的情况下得到广泛使用。但是，这种类型的应用程序面临的挑战之一是量化颜色信息，尤其是当颜色变化和阴影变化同时发生时。在这项工作中，以纸条的pH感测为例，开发了一种基于图像分析的颜色定量方法，该纸条的颜色和阴影都随pH值的变化而变化，并且容易受到环境光照的影响。通过在水溶液中拍摄pH试纸条获得图像。然后从图像中提取颜色信息，并在红色，绿色，蓝色（RGB）和色相，饱和度，值（HSV）颜色空间中探索其与pH的关系。发现HSV空间的色调与pH有良好的相关性，尽管感测条的阴影和颜色都发生了变化。然后通过将颜色转换为色相来量化感测条的颜色，并建立了预测水的pH值的反向模型。发现，模型的R ²为0.952，均方根误差（RMSE）为0.219，优于pH试纸的视觉解释精度（0.5）。最后，讨论了该模型在不同光照条件下的适用性。因此，已经实现了具有颜色和阴影变化的颜色量化。我们相信，这可以在遇到颜色和阴影变化的各种使用点设备中找到广泛的应用，例如在室外场景中。