当前位置: 首页 > 文章 > 茶叶嫩芽机器视觉识别算法研究 农业装备与车辆工程 2020 (4) 34-36+45
Position: Home > Articles > Research on Visual Recognition Algorithm of Tea Shooting Machine Agricultural Equipment & Vehicle Engineering 2020 (4) 34-36+45

茶叶嫩芽机器视觉识别算法研究

作  者:
邵佩迪;吴明晖;季亚波;王彬;丁润冬
单  位:
上海工程技术大学机械与汽车工程学院
关键词:
茶叶嫩芽;视觉识别;双边滤波;图像分割
摘  要:
针对自主采茶机器人,研究了在茶园自然光环境下如何高效识别茶叶嫩芽。针对自然光条件下采集的茶叶图像含有大量噪声的情况,为了避免一些像素值变化剧烈的像素点,根据分析,最终选用双边滤波去噪算法,对茶叶原始图像进行平滑滤波的同时,还能有效保留图形的边缘等有用信息。采用一种新的基于颜色通道调换的算法来增大茶叶嫩芽和老叶以及环境的对比度,然后提取茶叶的颜色特征,进而分割提取出茶叶嫩芽。实验结果表明:基于颜色通道变换的算法具有高效稳定等优点,能够很好地识别茶叶嫩芽,可以满足自主采茶机器人对茶叶嫩芽识别的要求。该算法为后续自主采茶机器人的研发提供了技术支持。
译  名:
Research on Visual Recognition Algorithm of Tea Shooting Machine
作  者:
Shao Peidi;Wu Minghui;Ji Yabo;Wang Bin;Ding Rundong;School of Mechanical and Automotive Engineering, Shanghai University of Engineering Science;
关键词:
tea bud;;visual identity;;bilateral filtering;;image segmentation
摘  要:
For the automatic tea picking robot,tea buds were efficiently identified in natural light environment of tea garden.In view of the fact that the image of tea collected under natural light conditions contains a lot of noise,in order to avoid some pixels with sharp changes in pixel value,the bilateral filtering denoising algorithm is selected according to the analysis to smooth the original image of tea leaves while retaining the graphics effectively.At the same time,it can effectively retain useful information such as the edges of graphics.A new color channel-based algorithm is used to increase the contrast between tea buds and old leaves and the environment,extract the color characteristics of the tea leaves,and then segment and extract the tea buds.The experimental results show that the algorithm based on color channel transformation has the advantages of high efficiency and stability,and can well identify tea buds,which can meet the requirements of tea picking robots for tea bud recognition.The algorithm provides technical support for the development of subsequent autonomous teapicking robots.

相似文章

计量
文章访问数: 18
HTML全文浏览量: 0
PDF下载量: 0

所属期刊

推荐期刊