光调制解调技术在田间杂草光谱识别中的应用

Application of optical modulation and demodulation technology in weed spectral identification

为了实现快速、无损地从作物中识别杂草,研究了基于光谱分析技术的光谱传感器.以苗期冬油菜为研究对象,根据已提取的4个特征波长(590,710,750,940 nm),设计了一套光谱传感器原型试验系统.该试验系统在传统光谱传感器的基础上,运用光调制解调技术剔除检测结果中环境杂散光的干扰,包括试验设备(光信号调制设备和光电信号采集设备)和试验数据处理LabVIEW程序.用冬油菜菜叶对试验系统4个不同波段进行了验证试验,当外界环境杂散光对应直流分量变化幅度分别为10. 00%,6. 40%,1. 17%,1. 34%,22. 60%,38. 90%,56. 00%,59. 50%时,反射率表征值一直保持稳定,表明在外界环境杂散光缓慢变化和剧烈变化时,系统可以稳定测量被测样品的反射率.该系统验证了光调制解调技术能有效提高被测样品光谱信号的测量信噪比.

In order to identify weeds from crops rapidly and nondestructively,a spectral sensor based on spectral analysis technology was developed.Winter rape in seedling stage was served as study crop,and a prototype experimental system with the spectral sensor was designed according to four characteristic wavelengths(590,710,750 and 940 nm),which already had been extracted.On basis of the traditional spectral sensor,the system could eliminate the interference of the environmental stray light in detected results by using optical modulation and demodulation technology.This system included experimental equipment(optical signal modulation equipment and photoelectric signal acquisition equipment)and experimental data processing LabVIEW program.Validation experiments were performed on four different bands using winter rape leaves.When the DC component variation amplitudes,which were caused from the change in the intensity of external environmental stray light,were 10.00%,6.40%,1.17%,1.34%,22.60%,38.90%,56.00%and 59.50%,respectively,the reflectance always was stable.The experimental results show that the system can stably measure the reflectance of measured samples when the ambient light intensity varies either slowly or rapidly.The system confirms that the optical modulation and demodulation technology can effectively improve the signal-to-noise ratio of the measured samples.