干燥过程中玉米水分质量分数均匀度的高光谱图像无损检测

Non-Destructive Detection of Moisture Content Uniformity During the Drying Process of Maize by Hyperspectral Imaging Technology

水分含量均匀度是干燥过程一个重要指标,它是评价干燥食品质量和干燥工艺一个重要参数。作者以干燥过程中的玉米为研究对象,研究高光谱图像技术检测水分质量分数均匀度的方法。采用均值特征和标准差特征结合偏最小二乘法(PLS)建立预测模型;并用正交信号校正法对均值特征和标准差进行预处理。结果表明:均值特征和标准差进行预处理后所建立的模型效果较好,预测相关系数为0.839,预测均方根误差为1.74%,潜在变量的数目为2个。研究表明:高光谱图像技术可用于水分质量分数均匀度的直接无损检测。

Moisture content uniformity is one of the primary parameters in a drying process,which is important to evaluate the quality of dried-foods and the drying technique. The moisture content uniformity detected by the hyperspectral imaging technology was studied in the drying process of maize. A prediction model was developed by the mean and standard deviation features combined with the partial least squares(PLS),where orthogonal signal correction method was used as the preprocessing method. The results showed that the prediction model developed by the mean and standard deviation features after preprocessing achieved the optimal performance with the correlation coefficient of 0.839 and the root mean square error of 1.74%,while the latent variables was reduced to 2 variables. Therefore,the hyperspectral imaging technology could be used as a non-destructive detection of moisture content uniformity.