Search

For scanning electron microscopes with high resolution and a strong electric field, biomass materials under observation are prone to radiation damage from the electron beam. This results in blurred or non-viable images, which affect further observation of material microscopic morphology and characterization. Restoring blurred images to their original sharpness is still a challenging problem in image processing. Traditional methods can't effectively separate image context dependency and texture information, affect the effect of image enhancement and deblurring, and are prone to gradient disappearance during model training, resulting in great difficulty in model training. In this paper, we propose the use of an improved U-Net (U-shaped Convolutional Neural Network) to achieve image enhancement for biomass material characterization and restore blurred images to their original sharpness. The main work is as follows: use of depthwise separable convolution instead of standard convolution in U-Net to reduce model computation effort and parameters; embedding wavelet transform into the U-Net structure to separate image context and texture information, thereby improving image reconstruction quality; using dense multi-receptive field channel modules to extract image detail information, thereby better transmitting the image features and network gradients, and reduce the difficulty of training. The experiments show that the improved U-Net model proposed in this paper is suitable and effective for enhanced deblurring of biomass material characterization images. The PSNR (Peak Signal-to-noise Ratio) and SSIM (Structural Similarity) are enhanced as well. [ABSTRACT FROM AUTHOR]

The aeronautics and aerospace industries often require special-shaped parts made from lightweight materials with a constant resistance, such as filament winding composite elbows and tees. Filament winding patterns can be realized using numerically controlled filament winding machines. Herein, a 3-axis computer controlled filament winding machine is proposed to solve existing problems with winding of composite elbows such as inconsistent quality, low productivity, and high costs. In this study, a geodesic winding equation for the torus and non-geodesic winding equation for the cylindrical sections of the elbow are provided and the winding angle α' is optimized. Furthermore, the correspondence relationship between the cylindrical and torus section motion is derived. The winding pattern is optimized and tested using the proposed 3-axis filament-winding machine. The results show that the optimal winding pattern design can be easily calculated with a programmable multi-axis controller using a simple control program, and a consistent winding pattern can be achieved. This paper provides a low-cost manufacturing method for the filament winding of composite elbows with cylindrical ends of unequal lengths. [ABSTRACT FROM AUTHOR]

Extraction of watershed characteristics (stream position, stream length, watershed area, slope, and others) using a digital elevation model (DEM) is an important prerequisite of hydrologic analysis. Detecting these surface hydrologic features in flat regions using a DEM is a known problem. Using a known stream network to improve DEM performance has been done in a variety of methods; many methods can improve calculation of watershed area, stream position, and length, but there are strong impacts (e.g., slope) on subsequent terrain analysis from altering the DEM. To solve this problem, a new method, water depth gradient burning (WDGB), is proposed in this paper to extract watershed characteristics and subsequent terrain analysis in a flat region. In WDGB, the stream network map is used as an input dataset; a burning value at headwater, stream connection, and outlet nodes of the stream is defined, which is consistent with three parts: water depth correction, peak depth correction, and additional depth for gradient, ensuring that an upstream node elevation is higher than a downstream one. Other stream elevations are calculated by linear interpolation using the aforementioned node elevations. Because the adjusted DEM (the stream) has a positive gradient with no pits or inverse slope along the flow direction, it can be used to extract watershed characteristics and obtain the hydrologically correct result. At the same time, the elevation of the stream is not altered too much in WDGB. The effect of the adjusted DEM on subsequent terrain analysis is addressed with regard to the D8 algorithm, AGREE, and WDGB methods. The result indicates that WDGB can not only achieve watershed characteristics correctly in a flat region but also have the least influence on the original DEM and the subsequent terrain analysis. [ABSTRACT FROM AUTHOR]