可溶性固形物含量（soluble solids content，SSC）是衡量水果内部质量的重要指标，为此，提出一种基于高光谱图谱融合的无损检测方法，用于预测蓝莓的SSC。采用3种典型的波长降维算法，包括蒙特卡罗无信息变量消除（monte carlo uninformative variable elimination，MC-UVE）、竞争性自适应重加权采样（competitive adaptive reweighted sampling，CARS）和连续投影算法（successive projections algorithm，SPA），用于筛选有效波长。此外，提出一种结合局部二值模式（local binary patterns，LBP）和灰度共生矩阵（gray level co-occurrence matrix，GLCM）提取图像特征的策略。基于光谱特征、图像特征和融合特征，分别建立偏最小二乘（partial least squares，PLS）、反向传播神经网络（back propagation neural network，BPNN）和支持向量机（support vector machine，SVM）模型进行SSC预测。研究结果表明，利用CARS算法提取的光谱特征融合LBP+GLCM算法提取的图像特征建立的BPNN模型，具有最佳的预测精度。该模型的决定系数（R ²）为0.926 1，均方根误差（root mean square error of prediction，RMSEP）为0.364 1。该研究表明高光谱图谱融合技术在无损预测蓝莓SSC中具有较大应用潜力。

Soluble solids content （SSC） is a key indicator for assessing the internal quality of fruits. This study proposes a non-destructive detection method based on hyperspectral image fusion to predict the SSC of blueberries. Three widely used wavelength dimensionality reduction algorithms are employed：Monte Carlo uninformative variable elimination （MC-UVE）， Competitive Adaptive Reweighted Sampling （CARS）， and Successive Projections Algorithm （SPA）， to identify optimal wavelengths. Additionally， a strategy integrating Local Binary Patterns （LBP） and Gray Level Co-occurrence Matrix （GLCM） is proposed for feature extraction. Using spectral features， image features， and fused features， Partial Least Squares （PLS）， Backpropagation Neural Network （BPNN）， and Support Vector Machine （SVM） models are developed for SSC prediction. The results demonstrate that the BPNN model， utilizing spectral features extracted via the CARS algorithm and image features derived from the LBP+GLCM algorithm， yields the highest prediction accuracy. The model's coefficient of determination （R {}_p^{\mathrm{2}}） is 0.926 1， while the Root Mean Square Error of Prediction （RMSEP） is 0.364 1. This study indicates that hyperspectral image fusion technology holds significant potential for the non-destructive prediction of blueberry SSC.