Grape quality inspection needs the identification of multiple categories of grapes, and there are many scenes such as depth of field changes and multiple strings in the grape images. Grape recognition is ineffective due to the limitations of single pretreatment method. The research objects were 15 kinds of natural scene grape images collected in the greenhouse, and the corresponding image dataset Vitis-15 was established. Aiming at the large intra-class differences and small inter-class of differences grape images, a multi-scale grape image recognition method based on Convolutional Neural Network (CNN) was proposed. Firstly, the data in Vitis-15 dataset were pre-processed by three methods, including the image rotating based data augmentation method, central cropping based multi-scale image method and data fusion method of the above two. Then, transfer learning method and convolution neural network method were adopted to realiize the classification and recognition. The Inception V3 network model pre-trained on ImageNet was selected for transfer learning, and three types of models-AlexNet, ResNet and Inception V3 were selected for convolution neural network. The multi-scale image data fusion classification model MS-EAlexNet was proposed, which was suitable for Vitis-15. Experimental results show that with the same learning rate on the same test dataset, compared with the augmentation and multi-scale image method, the data fusion method improves nearly 1% testing accuracy on MS-EAlexNet model with 99.92% accuracy, meanwhile the proposed method has higher efficiency in classifying small sample datasets.

Plane Wave Pseudopotential (PWP) Density Functional Theory (DFT) calculation is the most widely used method for material calculation. The projector calculation plays an important part in PWP-DFT calculation for the self-consistent iteration solution, while it often becomes a hinder to the speed-up of software. Therefore, according to the features of Graphic Processing Unit (GPU), a speed-up algorithm was proposed: 1) using a new parallel mechanism to solve the potential energy of nonlocal projector, 2) redesigning the distribution structure of data, 3) reducing the use of computer memory, 4) Proposing a solution to the related data problems of the algorithm. Eventually got 18-57 times acceleration, and reached the 12 seconds per step of the molecular dynamics simulation. In this paper, the testing time of running this model on GPU platform was analysed in detail, meanwhile the calculation bottleneck of the implementation of this method into GPU clusters was discussed

A real-valued extended propagator method algorithm based on virtual interpolated array named VIA-EPM real-valued algorithm was proposed to solve the utilization of Virtual Interpolated Array (VIA) in the non-circular Direction Of Arrival (DOA) algorithm. The real array output was virtually transformed by utilizing transformation matrix which was obtained through the real array manifold and virtual array manifold. The real part and imaginary part of the transformed array output were obtained, which could be reconstructed in series to extend dimensions according to the characteristic that the signal sources are real-valued, and then a Propagator Method (PM) DOA estimation algorithm was obtained after splitting the extended array output matrix. The simulation results show that, if sensor position errors exist, the performance of the new algorithm is similar to the calibrated extended propagator method real-valued algorithm (EPM real-valued algorithm) by using VIA for the calibrated sensor position data, and this algorithm also keeps the performance in array extension, high accuracy and high resolution, and the performance of the new algorithm is obviously better than the non-calibrated EPM real-valued algorithm in the case of two-dimension sensor position errors. Analysis of the computational complexity of VIA-EPM real-valued algorithm concludes that the new algorithm has the advantage of virtual interpolated array and non-circular characteristic, and its computational complexity is much lower than complex-valued algorithm.