Recently, the automatic detection of decayed blueberries is still a challenge in food industry. Early decay of blueberries happens on surface peel, which may adopt the feasibility of hyperspectral imaging mode to detect decayed region of blueberries. An improved deep residual 3D convolutional neural network (3D-CNN) framework is proposed for hyperspectral images classification so as to realize fast training, classification, and parameter optimization. Rich spectral and spatial features can be rapidly extracted from samples of complete hyperspectral images using our proposed network. This combines the tree structured Parzen estimator (TPE) adaptively and selects the super parameters to optimize the network performance. In addition, aiming at the problem of few samples, this paper proposes a novel strategy to enhance the hyperspectral image sample data, which can improve the training effect. Experimental results on the standard hyperspectral blueberry datasets show that the proposed framework improves the classification accuracy compared with AlexNet and GoogleNet. In addition, our proposed network reduces the number of parameters by half and the training time by about 10%.

中文翻译：

---

基于改进的深度残差3D卷积神经网络的高光谱图像蓝莓早衰检测与分类

近年来，蓝莓腐烂的自动检测仍然是食品行业的一大挑战。蓝莓的早期腐烂发生在表皮果皮上，这可能是利用高光谱成像模式检测蓝莓腐烂区域的可行性。提出了一种改进的深度残差 3D 卷积神经网络 (3D-CNN) 框架用于高光谱图像分类，以实现快速训练、分类和参数优化。使用我们提出的网络，可以从完整的高光谱图像样本中快速提取丰富的光谱和空间特征。这自适应地结合了树结构的 Parzen 估计器（TPE）并选择超参数来优化网络性能。另外，针对样本少的问题，本文提出了一种新的增强高光谱图像样本数据的策略，可以提高训练效果。在标准高光谱蓝莓数据集上的实验结果表明，与 AlexNet 和 GoogleNet 相比，所提出的框架提高了分类精度。此外，我们提出的网络将参数数量减少了一半，训练时间减少了约 10%。