Instance segmentation of fruit tree canopies from images acquired by unmanned aerial vehicles (UAVs) is significance for the precise management orchards. Although deep learning methods have been widely used in fields feature extraction and classification, there are still phenomena complex data strong dependence on software performances. This paper proposes a learning-based instance method litchi trees, which has simple structure lower requirements form. Considering that models require large...

Clouds and accompanying shadows, which exist in optical remote sensing images with high possibility, can degrade or even completely occlude certain ground-cover information images, limiting their applicabilities for Earth observation, change detection, land-cover classification. In this paper, we aim to deal cloud contamination problems the objective of generating cloud-removed images. Inspired by low-rank representation together sparsity constraints, propose a coarse-to-fine framework...

Due to the inevitable existence of clouds and their shadows in optical remote sensing images, certain ground-cover information is degraded or even appears be missing, which limits analysis utilization. Thus, cloud removal great importance facilitate downstream applications. Motivated by sparse representation techniques have obtained a stunning performance variety applications, including target detection, anomaly so on; we propose two-pass robust principal component (RPCA) framework for...

AIE supramolecular fluorescence sensor with tripod structure was designed and used to special recognize PA.

Light detection and ranging (LiDAR) sensors have been widely deployed on intelligent systems such as unmanned ground vehicles (UGVs) aerial (UAVs) to perform localization, obstacle detection, navigation tasks. Thus, research into range data processing with competitive performance in terms of both accuracy efficiency has attracted increasing attention. Sparse coding revolutionized signal led state-of-the-art a variety applications. However, dictionary learning, which plays the central role...

Cavitation in water turbine can induce blade erosion, efficiency decrease and unit vibration, some extreme conditions the reverse hammer caused by draft-tube cavity collapse may lead to unit-lifting accident. The purpose of this work was analyse influences cavitation on pressure pulsation characteristics a prototype pump-turbine. We conducted 3D CFD simulations giving different coefficients that volumes at inlet. In one severe conditions, propagation pressures across whole flow channel...

Abstract. Band registration is one of the most critical steps in production multi/hyperspectral images and determines accuracy applications directly. Because characteristics imaging devices some satellites, there may be a time difference between bands during push-broom imaging, which leads to displacements moving clouds with respect ground. And large number feature points gather around cloud contours due high contrast rich texture, resulting building transformation more suitable for making...

Abstract In the paper, we consider problem of two-dimensional (2D) phase retrieval, which recovers a 2D complex-valued wave field from magnitudes both and its Fourier transform. Due to absence measurements, prior information on is needed in order recover phase, feasible when phases are sparse. this improve retrieval accuracy by incorporating sparse constraint field. As sequel previous iterative projection approaches, approaches with realized based ‘soft thresholding’. It has superior...

During the runaway process of high-head turbine-generator units due to wicket gate failure, ball valve should be closed promptly prevent accidents, and dynamic hydraulic characteristics during closing studied. In this paper, flows through a prototype are simulated by latest 3D CFD method. First, steady in different openings calculated, we find that headloss coefficients agree well with experimental results. Then, water hammer led linear is both 1D Method Characteristic (MOC), reasonable...

Deep learning has been used to image compressive sensing (CS) for enhanced reconstruction performance. However, most existing deep methods train different models subsampling ratios, which brings additional hardware burden. In this paper, we develop a general framework named scalable (SDCS) the sampling and (SSR) of all end-to-end-trained models. proposed way, images are measured initialized linearly. Two masks introduced flexibly control ratios in reconstruction, respectively. To make model...

This paper considers solving the unconstrained $\ell_q$-norm ($0\leq q<1$) regularized least squares ($\ell_q$-LS) problem for recovering sparse signals in compressive sensing. We propose two highly efficient first-order algorithms via incorporating proximity operator nonconvex functions into fast iterative shrinkage/thresholding (FISTA) and alternative direction method of multipliers (ADMM) frameworks, respectively. Furthermore, $\ell_q$-LS problem, a sequential minimization strategy is...

To research the brick masonry wall between windows under low cyclic load test. Combining with ABAQUS finite element software simulated windows. Analyze destruction process, ductility and seismic performance of Explore beneficial influence on which pressure. The results study show that occurred flexural failure.With increase pressure, Brick will be decreased. But yield load, maximum bearing capacity limit is improved obviously. When compared 0.3MPa,the 0.4MPa, 0.5MPa, 0.6MPa increased by 8.3%...

Compressive sensing (CS) is a technique for estimating sparse signal from the random measurements and measurement matrix. Traditional recovery methods have seriously degeneration with matrix uncertainty (MMU). Here MMU modeled as bounded additive error. An anti-uncertainty constraint in form of mixed L2 L1 norm deduced model MMU. Then we combine to get an operator. Numerical simulations demonstrate that proposed operator has better reconstructing performance than traditional methods.

This work addresses the robust reconstruction problem of a sparse signal from compressed measurements. We propose formulation for which employs $\ell_1$-norm as loss function residual error and utilizes generalized nonconvex penalty sparsity inducing. The $\ell_1$-loss is less sensitive to outliers in measurements than popular $\ell_2$-loss, while has capability ameliorating bias convex LASSO thus can yield more accurate recovery. To solve this nonsmooth minimization efficiently, we...

Hyperspectral image (HSI) has some advantages over natural for various applications due to the extra spectral information. During acquisition, it is often contaminated by severe noises including Gaussian noise, impulse deadlines, and stripes. The quality degeneration would badly effect applications. In this paper, we present a HSI restoration method named smooth robust low rank tensor recovery. Specifically, propose structural decomposition in accordance with linear mixture model of HSI. It...

Abstract. Cloud detection is a vital preprocessing step for remote sensing image applications, which has been widely studied through Convolutional Neural Networks (CNNs) in recent years. However, the available CNN-based works only extract local/non-local features by stacked convolution and pooling layers, ignoring global contextual information of input scenes. In this paper, novel segmentation-based network proposed cloud images. We add multi-class classification branch to U-shaped semantic...