No abstract

This article proposes a novel hierarchical residual network with attention mechanism (HResNetAM) for hyperspectral image spectral-spatial classification to improve the performance of conventional deep learning networks. The straightforward convolutional neural network based models have limitations in exploiting the multi-scale spatial and spectral features, and this is the key factor in dealing with the high-dimensional nonlinear characteristics present in hyperspectral images. The proposed hierarchical residual network can extract multi-scale spatial and spectral features at a granular level, so the receptive fields range of this network will be increased, which can enhance the feature representation ability of the model. Besides, we utilize the attention mechanism to set adaptive weights for spatial and spectral features of different scales, and this can further improve the discriminative ability of extracted features. Furthermore, the double branch structure is also exploited to extract spectral and spatial features with corresponding convolution kernels in parallel, and the extracted spatial and spectral features of multiple scales are fused for hyperspectral image classification. Four benchmark hyperspectral datasets collected by different sensors and at different acquisition time are employed for classification experiments, and comparative results reveal that the proposed method has competitive advantages in terms of classification performance when compared with other state-ofthe-art deep learning models.

No abstract

It is desirable to fabricate materials with adjustable physical properties that can be used in different industrial applications. Since the property of a material is highly dependent on its inner structure, the understanding of structure–property correlation is critical to the design of engineering materials. 3D printing appears as a mature method to effectively produce micro-structured materials. In this work, we created different stainless-steel microstructures by adjusting the speed of 3D printing and studied the relationship between thermal property and printing speed. Our microstructure study demonstrates that highly porous structures appear at higher speeds, and there is a nearly linear relationship between porosity and printing speed. The thermal conductivity of samples fabricated by different printing speeds is characterized. Then, the correlation between porosity, thermal conductivity, and scanning speed is established. Based on this correlation, the thermal conductivity of a sample can be predicted from its printing speed. We fabricated a new sample at a different speed, and the thermal conductivity measurement agrees well with the value predicted from the correlation. To explore thermal transport physics, the effects of pore structure and temperature on the thermal performance of the printed block are also studied. Our work demonstrates that the combination of the 3D printing technique and the printing speed control can regulate the thermophysical properties of materials.

No abstract