提出了基于高光谱成像技术的猪肉嫩度检测方法。利用高光谱成像系统获取78个猪肉样本在400~1100 nm范围的高光谱图像数据；通过主成分分析高光谱数据进行降维，从中优选出3幅特征图像，并从每幅特征图像中分别提取对比度、相关性、角二阶矩和一致性等4个基于灰度共生矩阵的纹理特征变量，这样每个样本共有12个特征变量，再通过主成分分析提取6个主成分变量，并参照剪切力方法测得的样本嫩度等级结果，利用神经网络方法构建猪肉嫩度等级判别模型。模型对校正集样本的回判率为96.15%，预测集样本的判别率为80.77%。研究表明高光谱图像技术可以用于猪肉嫩度等级水平的检测。

Detection of pork tenderness by hyperspectral imaging technique was proposed. First, hyperspectral images of 78 pork samples were captured by hyperspectral imaging system, and the spectral region is from 400 to 1100 nm. Dimension reduction was implemented on hyperspectral data by principal component analysis (PCA) to select 3 characteristic images. Next, 4 characteristic variables were extracted by texture analysis based on gray level cooccurrence matrix (GLCM), and they are contrast, correlation, angular second moment, and homogeneity, respectively, thus 12 characteristic variables in total for 3 characteristic images. PCA was conducted on 12 characteristic variables, and 6 principal component variables were extracted as the input of the discrimination model. The detection model of pork tenderness was constructed by artificial neural network (ANN), according to the reference results of pork tenderness by WarnerBratzler method. Detection results of ANN model are 96.15% and 80.77% in calibration and prediction sets, respectively. This work shows that it is feasible to detect pork tenderness by hyperspectral imaging technique.