基于形态学图像检测的机械手移栽穴苗识别技术

Recognition Technology of Transplanting Seedling of Robot Hand Based on Morphological Image Detection

在机械手移栽穴苗的过程中,针对取钵体导致的苗体根系存在不同程度的残缺、外漏等问题,提出一种基于形态学图像检测的机械手移栽穴苗识别技术。该技术分为以下3步:(1)通过确定形态学图像处理方法的衡量指标,确定了研究所采用的方案;(2)采用基于形态学的方法对穴苗进行图像处理仿真实验;(3)开发取钵体效果判定的软件系统并进行测试和实验。采用西瓜幼苗进行仿真实验,仿真结果表明:基于形态学图像处理的机械手移栽穴苗识别技术具有较好的实时性、稳定性和鲁棒性。

In the process of mechanical transplanting seedling in hand because seedling roots from the pot body lead to the existence of incomplete and leakage problems in different degree,this paper presents a morphological image detection of mechanical transplanting seedling identification technology based on hand. This technology is divided into the following three steps：（1）by determining the index of morphological image detection method,so as to determine the research on the scheme;（2）using the method based on morphology of seedling image processing simulation experiment;（3） the development effect of bowl judgment system and test and experiment. The simulation results show that the recognition technique based on morphological image detection has good real-time performance,stability and robustness.