Study on detection of moisture content in lettuce leaves based on hyperspectral imaging technology___ | 基于高光谱成像技术的生菜叶片水分检测研究_____

Китайский. ［目的］ 探索利用高光谱图像技术检测作物含水率的方法。［方法］ 以意大利全年耐抽苔生菜为试材，利用高光谱成像系统采集生菜叶片的高光谱图像，用ENVI V.4和Matlab V.7.0软件对高光谱图像进行处理。［结果］ 采用自适应波段选择法从所采集的生菜叶片高光谱图像数据中优选出特征波长1 420 nm；对每个样本特征波长下的图像进行分割，反转以及形态运算等操作得到目标图像；从每个目标图像中提取灰度均值、灰度标准差作为灰度特征，能量、熵、惯性矩、相关性的均值和标准差作为纹理特征；采用GA-PLS法选出最优特征子集，并建立基于最优特征的偏最小二乘回归模型，以检测生菜叶片的含水率。［结论］ 模型的预测值与实测值的相关系数_R_为0.902，精度明显高于基于灰度特征或纹理特征的预测模型。

Английский. ［Objective］ The study aimed to explore the method for detecting the moisture content of the crop by hyperspectral imaging technology. ［Method］ With Italy lettuce with the bolting resistance in whole year as the tested materials, the hyperspectral images of lettuce leaves were collected by the hyperspectral imaging system and treated by ENVI V.4 and Matlab V 7.0 software. ［Result］ The optimal characteristic wavelength was 1 420 nm optimized from the selected hyperspectral images of lettuce leaves by using the adaptive band selection. The images of all samples at 1 420 nm were segmented, reversed and operated and then the target images were obtained. From each target image, the mean value and standard deviation of the gray scale were extracted as the gray feature, and the mean value and standard deviation of energy, entropy, moment of inertia and correlation were extracted as the texture feature. The optimal feature subset was selected by GA-PLS, and the partial least-squares regression model was established on base of the optimal characteristic to detecting the moisture content of the lettuce leaves. ［Conclusion］ The correlation coefficient between the predict value and the real value was 0.902, whose precision was obviously higher than the prediction models based on gray or texture feature._