With booming of computer technology, it is feasible to quantitatively extract valuable information from readily available image of objects, which dramatically reduces the cost and improves the efficiency. In this work, a structural classification method, based on data extracted from color formation process on thin layer chromatography (TLC) plates through computer processing, was established for natural products. Representative natural products exhibited good clustering and separation according to the extracted color feature, and structure of natural products can be classified by the distribution region in the color system. This method provides structural information for typical types of natural products directly from the formed color on TLC, which is very efficient and make portable device-based structure analysis of natural products possible. In addition, a general mechanism of color formation was proposed. This method is free from special instrument, high-throughput, and would facilitate large-scale screening of bioactives from natural sources.

随着计算机技术的蓬勃发展，从易于获得的物体图像中定量提取有价值的信息是可行的，这大大降低了成本并提高了效率。在这项工作中，建立了一种基于天然产品的薄板色谱（TLC）板颜色形成过程中提取的数据，并通过计算机处理的结构分类方法。代表性的天然产物根据提取的颜色特征表现出良好的聚类和分离，并且可以通过颜色系统中的分布区域对天然产物的结构进行分类。此方法直接从TLC上形成的颜色直接提供典型天然产品类型的结构信息，这非常有效，并使基于便携式设备的天然产品结构分析成为可能。另外，提出了形成颜色的一般机理。该方法无特殊仪器，高通量，将有助于大规模筛选天然来源的生物活性物质。