基于计算机视觉的茶叶色泽检测研究

我国茶叶种植面积和产量均为世界第一,是特色农业的重要组成部分;但我国茶叶品质检测体系不完善,分级技术水平不高,影响了产品在国际市场上的竞争力。传统的茶叶分级是由人工分析判断,具有较大的局限性。计算机视觉是一种新型的图像处理技术,已经应用于茶叶品质分析。为此,将拍摄的茶叶和茶水图像进行预处理、灰度化和阈值分割,获得目标轮廓并分析颜色特征,并通过建模集样本确定用于色泽检测的特征量,然后对检验集样本进行色泽检测。结果表明:检验集中被错误识别的茶叶种类极少,总体的识别准确率达到9 0%,为准确评价茶叶的色泽品质提供了技术支持。

Research on Tea Color Detection Based on Computer Vision

The planting area and output of tea in our country are both the highest in the world and an important part of characteristic agriculture. However, China's tea quality testing system is not perfect, grading technology is not high, affecting the product's competitiveness in the international market. Traditional tea classification is judged by manual analysis, with great limitations. Computer vision is a new type of image processing technology that has been applied to tea quality analysis. In this paper, the images of tea and tea are taken for preprocessing, graying and threshold segmentation to obtain the target contour and analyze the color features. Determine the amount of features used for color detection by modeling a set of samples, and then test the color of the test sample set. The types of tea that were misidentified in the test were very few, and the overall recognition accuracy reached 90%. This method of computer vision analysis can be used to detect the color of tea and provide technical support for accurately evaluating the color quality of tea.