利用高光谱成像技术和多变量校正方法检测苹果的硬度（英）

高光谱图像集图像信息与光谱信息于一身，应用于农产品品质无损检测领域。该研究尝试利用高光谱图像技术结合多变量校正方法检测苹果硬度的可行性。试验通过获取的高光谱图像中提取有效的光谱信息来建立预测苹果硬度的预测模型。在建立模型过程中，偏最小二乘（PLS）和支持向量回归（SVR）两种多变量校正方法被比较，结果表明在785.11～872.45 nm范围内，SVR模型的性能优于PLS模型，模型对硬度预测结果的相关系数为0.6808。试验结果表明高光谱图像技术可以被用来检测苹果的硬度。

Determination of apple firmness using hyperspectral imaging technique and multivariate calibrations

Hyperspectral imaging technology is applied to nondestructive quality determination of agricultural and food products. It has a greater advantage of combining spatial image and spectral measurement which can determine both external and internal quality of the product. Feasibility of using hyperspectral imaging technique and multivariate calibrations to determine apple firmness was studied. Forecasting model of apple firmness was established by effective spectral information extracted in hyperspectral image. Support vector regression (SVR) and partial least square (PLS) were applied comparatively to calibrate model. The result showed that the optimal spectral range of apple firmness was 785.11-872.45 nm. The SVR calibration model was superior to PLS model in fruit firmness determination. The correlation coefficient between the hyperspectral imaging prediction results and reference measurement results was R=0.6808 in the prediction. In conclusion, hyperspectral imaging technique can be applied to determine apple firmness.