茶叶的计算机识别应用研究

Research on Application of Computer Vision in the Identification of Tea

探讨了茶叶颜色实时检测的计算机系统的硬件组成以及有效的颜色检测模型和识别算法,并通过实验对其有效性进行了验证。探索了以茶的HIS图像为模型,提取H作为识别指标的遗传神经网络新途径,对茶叶图像进行了有效的识别。用所建立的识别指标和方法对不同烘炒工艺的茶叶进行检测,检测结果与人工判别结果的吻合率为91.7%,说明用计算机视觉代替人工感官进行识别是可行的。该方法的检测为茶叶感官品质检测提供了一种新的检测方法,应用前景广阔。

A computer vision algorithm based on color characteristic and colorful image was developed to identify tea leaf. To identify tea leaf, the characteristiic parameter of tea HIS model was extracted based on computer and Genetical-Neuralnetwork after the image conversion and preprocess, The experiments revealed that the method improved the consistence between computer inspection and manual inspection.