夜间自然环境下荔枝采摘机器人识别技术

发布时间：2018-01-27 00:33

本文关键词： 荔枝采摘机器人识别夜间图像 Hough圆检测　出处：《农业机械学报》2017年11期 　论文类型：期刊论文

【摘要】：利用机器视觉实现自然环境下成熟荔枝的识别,对农业采摘机器人的研究与发展具有重要意义。本文首先设计了夜间图像采集的视觉系统,然后选取了白天和夜间两种自然环境下采集荔枝图像,分析了同一串荔枝在白天自然光照与夜间LED光照下的颜色数据,确定了YIQ颜色模型进行夜间荔枝果实识别的可行性。首先选择夜间荔枝图像的I分量图,利用Otsu算法分割图像去除背景,然后使用模糊C均值聚类算法分割果实和果梗图像,得到荔枝果实图像;再利用Hough圆拟合方法检测出图像中的各个荔枝果实。荔枝识别试验结果表明:夜间荔枝图像识别的正确率为95.3%,识别算法运行的平均时间为0.46 s。研究表明,该算法对夜间荔枝的识别有较好的准确性和实时性,为荔枝采摘机器人的视觉定位方法提供了技术支持。
[Abstract]:The recognition of mature litchi in natural environment by machine vision is of great significance to the research and development of agricultural harvesting robot. Firstly, the visual system of night image acquisition is designed in this paper. Then the litchi images were collected in the daytime and nighttime environment, and the color data of the same litchi under daytime and LED illumination were analyzed. The feasibility of using YIQ color model to identify litchi fruit at night was determined. Firstly, I component map of nighttime litchi image was selected, and Otsu algorithm was used to segment the image to remove background. Then the fruit and stem images were segmented by fuzzy C-means clustering algorithm, and the litchi fruit image was obtained. The Hough circle fitting method was used to detect each litchi fruit in the image. The result of litchi recognition experiment showed that the accuracy rate of night litchi image recognition was 95.3%. The average operating time of the recognition algorithm is 0.46 s. The research shows that the algorithm is accurate and real-time for the identification of litchi at night, and provides technical support for the visual positioning method of litchi picking robot.
【作者单位】：华南农业大学数学与信息学院;
【基金】：国家自然科学基金项目(31201135、31571568) 广东省科技计划项目(2015A020209123) 广州市科技计划项目(201506010081)
【分类号】：TP242;TP391.41
【正文快照】： 引言果蔬采摘机器人已经逐步深入到农业生产[1]。开发具有视觉功能的智能果蔬采摘机器人,对于提高农业生产力有极大的应用价值与现实意义[2]。采摘机器人的工作环境为非结构化的自然环境,在自然环境中存在着各种随机因素,因此实现自然环境下采摘机器人对目标的视觉精确定位有

【相似文献】