冬油菜田杂草探测光谱传感器设计与应用

发布时间：2018-06-11 13:03

  本文选题：传感器 + 设计　； 参考：《农业工程学报》2017年18期

【摘要】：杂草的精确识别是对靶施药和自动化机械除草的关键前提,基于光谱分析技术的光谱传感器可以实现快速、无损的杂草识别。该文以冬油菜苗期杂草为研究对象,根据试验选取的4个特征波长点(595、710、755和950 nm),设计了一种能自动识别杂草的光谱传感器。根据光学系统原理和田间实际操作要求,提出了该光谱传感器的结构设计方案,选择了合适的光学器件,并开发了光谱传感器信号调理电路。对光谱传感器进行了标定和试验验证,根据便携式光谱仪和光谱传感器在4个波长下的测量结果建立了相应的标定方程,方程的决定系数分别为0.799、0.812、0.892和0.867,验证试验结果的相对误差绝大多数都在10%以内,可以识别冬油菜苗期田间杂草。该传感器为杂草自动探测装置的开发提供了参考。
[Abstract]:The accurate identification of weeds is the key premise of target application and automatic mechanical weeding. The spectral sensor based on spectral analysis technology can realize fast and lossless weed identification. Based on the four characteristic wavelengths selected in the experiment, a spectral sensor was designed to identify weeds in winter rape seedling stage according to the four characteristic wavelength points (595710755 and 950 nm). According to the principle of the optical system and the practical operation requirements in the field, the structural design of the spectral sensor is proposed, the appropriate optical device is selected, and the signal conditioning circuit of the spectrum sensor is developed. The calibration and experimental verification of the spectrum sensor are carried out, and the calibration equations are established according to the measurement results of the portable spectrometer and the spectrum sensor at four wavelengths. The determination coefficients of the equation were 0.799 ~ 0.812 ~ 0.892 and 0.867, respectively. The relative error of the test results was mostly less than 10%, which could be used to identify the weeds in winter rape seedling stage. The sensor provides a reference for the development of automatic weed detection device.
【作者单位】： 江苏大学现代农业装备与技术教育部重点实验室;
【基金】：国家自然科学基金项目(51575244) 江苏省高校自然科学研究项目(14KJA210001) 公益性行业(农业)科研专项经费项目(201503130) 江苏高校优势学科建设工程资助项目(2014-37) 江苏大学高级人才基金资助项目(14JDG149)
【分类号】：S451;S565.4;TP212
，

本文编号：2005343