Authors

Abstract

In the scene of paddy field rotary tillage, a real-time detection method of rotary tillage condition based on machine vision is proposed, and the quality of rotary tillage is evaluated by the index of residual stubble. The residual root stubble is selected as the research object, and the root stubble detection method based on the standard deviation of Y component in YCrCb space is proposed to determine the residual root stubble of soil after rotary tillage, which is divided into three levels: less root stubble, medium root stubble, and more root stubble. Finally, the accuracy of the algorithm is verified by field test and questionnaire survey. On the basis of manual evaluation, the accuracy rate of the working condition is 83.6 %, which provides a more accurate basis for the real-time adjustment of the control strategy for the unmanned operation of agricultural machinery in the field, and realizes the rotary tillage quality from qualitative evaluation to quantitative evaluation, and lays the foundation for the data of rotary tillage quality.

Abstract in Chinese

在面向水田旋耕作业场景中，提出基于机器视觉的旋耕工况实时检测方法，以残余根茬量的指标来评价旋耕质量。选取残余根茬量为研究对象，提出基于YCrCb空间Y分量的标准差的根茬量检测方法来判断旋耕后土壤的残余根茬量，分为三个等级：根茬量较少、根茬量中等、根茬量较多；最终以现场试验和问卷调查的形式验证算法的准确性。在人工评判的基础上，该工况的准确率为83.6%，这对大田农机无人化作业的控制策略的实时调整提供较为准确的依据，实现对旋耕质量从定性评价到定量评价的转换，为旋耕质量数据化奠定基础。