Accurate landslide extraction is significant for disaster prevention and control. Remote sensing images have been widely used in investigation, methods based on deep learning combined with remote (such as U-Net) received a lot of attention. However, because the variable shape texture features landslides images, rich spectral features, complexity their surrounding using U-Net can lead to problems such false detection missed detection. Therefore, this study introduces channel attention...

In this paper, we propose a novel stock price prediction model based on deep learning. With the success of learning algorithms in field Artificial Neural Network (ANN), choose to solve regression problems (stock our case). Stock is challenging problem due its random movement. This hybrid combination two well-known networks, Long Short Term Memory (LSTM) and Gated Recurrent Unit (GRU). We S&P 500 historical time series data use significant evaluation metrics such as mean squared error,...

Remote sensing monitoring of glacial lakes is an indispensable tool for identifying and preventing lake disasters. At present, the existing extraction methods based on Landsat remote image have achieved remarkable results, but algorithms used lack ability to analyze spectral shape texture features, require manual design parameters fine tune automation algorithm. As a result, it cannot mine depth features glacier in images accurately enough. To address these challenges, this study designed...

Shanghai is susceptible to land subsidence due its unique geological environment and frequent human activities. Traditional leveling techniques are not sufficient for monitoring large areas of the time-consuming, labor-intensive, expensive nature process. Furthermore, results conventional methods may be timely, rendering them ineffective purposes. Interferometric Synthetic Aperture Radar (InSAR) technology a widely used method ground low cost, high efficiency, ability cover areas. To monitor...

We develop an integrated neural network landslide susceptibility assessment (LSA) method that integrates temporal dynamic features of interferometry synthetic aperture radar (InSAR) deformation data and the spatial influencing factors. construct a time-distributed convolutional (TD-CNN) bidirectional gated recurrent unit (Bi-GRU) to better understand InSAR cumulative deformation, multi-scale (MSCNN) determine factors, parallel unified deep learning model fuse these for LSA. Compared with...

The surface deformation caused by underground mining leads to damage buildings and brings potential safety hazards property losses. demand for reliable prediction methods of in areas is becoming increasingly significant. At present, most are based on sampling points; however, these neglect consider local overall spatial features, this oversight affects the accuracy results. data form output often discontinuous not intuitive. In order solve problem, spatiotemporal method a very effective...

Landslide susceptibility mapping (LSM) is very important for hazard risk identification and prevention. Most of existing neural network models extract a pixel neighborhood feature or sequence landslide factors on one side, which leads to the generalization ability difficultly, had low prediction accuracy in complex scenes. In this paper, new unified information considering superimposed spatial neighbourhood proposed LSM. Different from traditional model framework, conditioning are merged...

Surface deformation poses a great threat to the safety of Jinchuan mining area production activities. At present, spatio-temporal evolution law and mechanism surface in are unclear, it is difficult obtain reliable prediction results using existing methods due lack model parameters or relevant data. To solve these problems, this study proposes new unified convolutional neural network with peephole long short-term memory (CNN-PhLSTM). Small baseline subset interferometric synthetic aperture...

Landslides are destructive geological hazards that occur all over the world. Due to periodic regulation of reservoir water level, a large number landslides in Three Gorges Reservoir area (TGRA). The main objective this study was explore preference machine learning models for landslide susceptibility mapping TGRA. Wushan segment TGRA selected as case study. At first, 165 were identified and total 14 causal factors constructed from different data sources. Multicollinearity analysis information...

Landslide susceptibility mapping (LSM) is of great significance in geohazard early warning and prevention. The existing LSM methods mostly used traditional static landslide conditioning factors (LCFs), which only considered the spatial features single-pixel neighborhoods could not extract time-series dynamic developing landslides, resulting low accuracy insufficient reliability LSM. To solve this problem, study proposes to introduce rainfall based on construct an integrated neural network...

Laryngoscopy, the most common diagnostic method for vocal cord lesions (VCLs), is based mainly on visual subjective inspection of otolaryngologists. This study aimed to establish a highly objective computer-aided VCLs diagnosis system deep convolutional neural network (DCNN) and transfer learning.To classify VCLs, our combined DCNN backbone with learning specifically finetuned laryngoscopy image dataset. Laryngoscopy database was collected train proposed system. The performance compared...

The speckle noise found in synthetic aperture radar (SAR) images severely affects the efficiency of image interpretation, retrieval and other applications. Thus, effective methods for despeckling SAR are required. traditional fail to balance terms relationship between intensity filtering retention texture details. Deep learning based have been shown potential achieve this balance. Therefore, study proposes a self-attention multi-scale convolution neural network (SAMSCNN) method despeckling....

Accurate and efficient landslide identification is an important basis for disaster prevention control. Due to the diversity of features, vegetation occlusion, complexity surrounding surface environment in remote sensing images, deep learning models (such as U-Net) detection based only on optical images will lead false missed detection. The accuracy not high, it difficult satisfy demand. SAR has penetrability, are highly sensitive changes morphology structure. In this study, a multi-input...

Background: Muscle infiltration of bladder cancer has become the most important index to evaluate its prognosis. Machine learning is expected accurately identify muscle status on images. Objective: This study aimed establish and validate a machine prediction model based multi-phase contrast-enhanced CT (MCECT) for preoperatively evaluating muscle-invasive cancer. Methods: A retrospective was conducted patients who underwent surgical resection pathological confirmation by MCECT scans. They...

The anomalous movements of glaciers cause disasters, such as debris flows and landslides. It is very important to assess the glacier their future trends. Glacier velocity refers movement process. current research aims analyse past spatiotemporal changes in velocity. No study has used neural network model conduct a prediction for Therefore, this paper selected typical mountain G2 G5 along Sichuan-Tibet Railway objects constructed Convolutional Gate Recurrent Unit (ConvGRU) based on 1988–2018...

The ancient Suoertou landslide seriously threatens the surrounding population's lives and property. Monitoring predicting this is crucial to ensure affected areas' safety. previous research on landslide's displacement characteristics mechanisms has lacked detailed analyses. Additionally, its future development trends must be understood. Therefore, we conducted a analysis of using small baseline subset interferometric synthetic aperture radar (SBAS-InSAR) monitoring results from 2018 2023....

Surface deformation of alpine peatland in China has an important effect on runoff and is great significance for wetland ecosystem protection. However, spatio-temporal characteristics surface lack systematic studies, the driving mechanism not yet clear. In this study, we selected Maduo County as research object, based small baseline subset radar interferometry (SBAS-InSAR) technique was obtained, analyzed patterns peatland, explored with single-factor multi-factor combinations Geo-detector,...

Abstract. The Poiqu River basin is an area of concentration for glaciers and glacial lakes in the central Himalayas, where 147 were identified, based on perennial remote sensing images, with lake ranging from 0.0002 to 5.5 km2 – a total 19.89 km2. Since 2004, retreat rate glacier has reached as high 5.0 a−1, while growth 0.24 a−1. We take five typical our case study find that reaches 31.2 %, expanded by 166 %. Moreover, we reconstruct topography calculate water capacity propose balance...