Hyperspectral image (HSI) change detection is a technique for detecting the changes between multitemporal HSIs of same scene. Many existing methods have achieved good results, but there still exist problems as follows: 1) mixed pixels in HSI due to low spatial resolution hyperspectral sensor and other external interference 2) many deep learning-based networks cannot make full use correlation difference information bitemporal images. These are not conducive further improving accuracy...

Recently, deep neural networks (DNNs) have been widely used in hyperspectral image change detection (HSI-CD). Generally, training such a DNN-based HSI-CD network often requires large number of labeled samples. However, it is time-consuming, labor-intensive, or even infeasible to label samples practice. In this article, we propose feature mutual representation-based graph domain adaptive (FGDANet) for unsupervised HSI-CD. This method constructs pseudosiamese backbone consisting two customized...

Hyperspectral image super-resolution (HISR) is a technique that can break through the limitation of imaging mechanism to obtain hyperspectral (HSI) with high spatial resolution. Although some progress has been achieved by existing methods, most them directly learn spatial-spectral joint mapping between observed images and target high-resolution HSI (HrHSI), failing fully reserve spectral distribution low-resolution (LrHSI) multispectral imagery (HrMSI). To this end, we propose...

Representative driving cycles and specific vehicle types play an important role in analyzing energy consumption to provide valuable references for manufacturers government inspections. In this study, a parallel plug-in hybrid electric with different driven configurations is employed analyze under typical cycle. Specifically, cycle of Tianjin constructed by integrating the clustering Markov chain algorithms. Considering various layouts systems vehicles, three (P2, P3, P4) are established....

Abstract The existing displacement of parallel + serial type hybrid mechanisms is mainly solved by the equivalent mechanism (SM) method. However, a large number lower mobility (PMs) that have high coupled motions at end-effector cannot be to SMs. Thus, problem especially for inverse this has not been well solved. On basis situation, article takes 6-degrees-of-freedom (DOFs) 3-UPU 3R as an example give general method solve problem. First, based on and pose coupling relationship PM, its...

Hyperspectral image (HSI) change detection is a technique to observe the information between multitemporal HSIs, which currently considered major focus of research in filed remote sensing intelligent interpretation. Most existing deep learning-based methods have created satisfactory performance, but these lack transparency and poor generalization. To tackle problems outlined above, we propose dictionary learning-guided interpretable network for hyperspectral detection, unfolds model into an...

Abstract Numerical study has been carried out to improve the unreasonable oil film accumulation and return effect of bearing chamber. Ramp sump eccentricity offtake structures are designed improved, oil-gas two-phase flow calculation model based on CLSVOF (coupled level set volume fluid) method is proposed. Based grid-independent analysis verifying rationality numerical data, movement mechanism characteristics for different scavenge offtakes calculated analyzed. Results show that both ramp...

More and more evidence has shown that strengthening layer interactions can enhance the representation power of a deep neural network, while self-attention excels at learning interdependencies by retrieving query-activated information. Motivated this, we devise cross-layer attention mechanism, called multi-head recurrent (MRLA), sends query current to all previous layers retrieve query-related information from different levels receptive fields. A light-weighted version MRLA is also proposed...

Optical diffractive neural networks have triggered extensive research with their low power consumption and high speed in image processing. In this work, we propose a reconfigurable digital all-optical network (R-ODNN) structure. The optical neurons are built Sb2Se3 phase-change material, making our reconfigurable, digital, non-volatile. Using three layers 14,400 on each 10 photodetectors connected to resistor network, model achieves 94.46% accuracy for handwritten digit recognition. We also...

In the construction of "Double High School", integrated teacher-student development has been internal requirements for establishing a team high-level double-position teachers and cultivating talents with high-quality technical skills. study, structured teaching is built individualized training "one teacher, one case" as starting point; promoted by creating "golden course", research interest, determining social service direction, setting up an innovation students; “one case” cultivation...

The reform of the domestic fire control system has brought about strong social demand for innovation in terms updating engineering education system. There are many existing problems training talents at local colleges, including unclear objectives, an unreasonable curricular system, overly standardized teaching method, neglect students' personal development, limited practical capability teachers, evaluation methods, and too great emphasis on learning results, etc.; all which relate to concept...

The technological innovation capability of the western region is key factor for coordinated development China's overall economy. Based on panel data 12 provinces in Western China from 2012 to 2018, this paper establishes evaluation index system ability China, and uses analysis get score ranking past six years, divides into positive type conservative backward type. This analyses results research, summarizes factors that restrict upgrading provinces, puts forward paths Countermeasures regions.

The clarity of seismic images has a great impact on the results data interpretation; therefore, high-resolution can provide finer geological information for exploration. Coordinate-based Multilayer Perceptron(MLP) is easy gradient-based optimization and machine learning, represents more compact order magnitude than grid sampling. In this paper, we propose combination Random Fourier Features(RFF) coordinate based MLP method to achieve super-resolution reconstruction images, taking one image...

This paper proposes a migration learning-based design model for an intelligent FH anti continuous change jamming system. To be able to withstand ongoing changes in linear sweep and random periodic jamming, this builds on the existing reinforcement learning frequency hopping anti-jamming method adds idea of it. The core point behind algorithm is that party will restart training system continue optimize current basis when it notices communication quality has decreased as result order maintain...

This paper presents a high altitude platform (HAP) communication system which applies the idea of satellite technology to it. The can offer extensive circular coverage on ground when base station and antenna are deployed high-altitude with fixed height 20 kilometers. uses theoretical concept transfer issues air equipment in order solve them because current HAP model will face significant challenges its actual implementation, be primarily focused side. By using merely circularly polarized...