A5051265102- Remote-Sensing Image Classification
- Remote Sensing and Land Use
- Advanced Image Fusion Techniques
- Infrared Target Detection Methodologies
- Metaheuristic Optimization Algorithms Research
- Remote Sensing in Agriculture
- Space Satellite Systems and Control
- Remote Sensing and LiDAR Applications
- Target Tracking and Data Fusion in Sensor Networks
- Planetary Science and Exploration
- Inertial Sensor and Navigation
- Automated Road and Building Extraction
- Advanced Semiconductor Detectors and Materials
- Calibration and Measurement Techniques
- Radiative Heat Transfer Studies
- Advanced Chemical Sensor Technologies
- Astro and Planetary Science
- Optical Systems and Laser Technology
Shanghai Institute of Technical Physics
2021-2023
University of Chinese Academy of Sciences
2021-2023
Chinese Academy of Sciences
2022-2023
Finnish Geospatial Research Institute
2021
Digital maps of road networks are a vital part digital cities and intelligent transportation. In this paper, we provide comprehensive review on extraction based various remote sensing data sources, including high-resolution images, hyperspectral synthetic aperture radar light detection ranging. This is divided into three parts. Part 1 provides an overview the existing acquisition techniques for extraction, methods, typical sensors, application status, prospects. 2 underlines main methods...
Dimensionality reduction (DR) is of great significance for simplifying and optimizing hyperspectral image (HSI) features. As a widely used DR method, kernel minimum noise fraction (KMNF) transformation preserves the high-order structures original data perfectly. However, conventional KMNF estimation (KMNF-NE) uses local regression residual neighbourhood pixels, which depends heavily on spatial information. Due to limited resolution, there are many mixed pixels in HSI, making KMNF-NE...
The mid-wave infrared imager has stronger transmission ability than the visible near-infrared imager, which can effectively overcome limitation of low-visibility climate on image acquisition timeliness. Moreover, it is more conducive to detection long-distance small target. However, few systems track aerial target while detecting in real-time. Meanwhile, conventional optical structures for imaging are cumbersome and bulky. Thus, significant challenge solve this problem how design a...
Aiming at the problem wherein temperature inversion accuracy is unstable due to major differences in atmospheric transmittance under various observation paths, a method for measuring radiation characteristics of an aircraft engine’s hot parts and skin using cooled middle-wave infrared camera proposed. Based on analysis aircraft’s characteristics, transmission model any path was revised, absolute correction established, equation calculated. Then, we used quasi-Newton calculate discussed...
Under the too short arc scenario, evolutionary-based algorithm has more potential than traditional methods in initial orbit determination. However, underlying multimodal phenomenon determination is ignored by current works. In this paper, we propose a new enhanced differential evolution (DE) with property to study angle-only IOD problem. Specifically, coarse-to-fine convergence detector implemented, based on Boltzmann Entropy, determine evolutionary phase of population, which lays basis...
Feature extraction, aiming to simplify and optimize data features, is a typical hyperspectral image dimensionality reduction technique. As kernel-based method, kernel minimum noise fraction (KMNF) transformation excellent at handling the nonlinear features within HSIs. It adopts function ensure linear separability by transforming original higher feature space, following which analysis can be performed in this space. However, KMNF has problem of high computational complexity low execution...
The airborne hyperspectral remote sensing systems (AHRSSs) acquire images with high spectral resolution, spatial and temporal dimension. While the AHRSS captures more detailed information from terrain objects, computational complexity of data processing is greatly increased. As an important application technology in domain, anomaly detection (AD) must be real-time high-precision many cases, such as post-disaster rescue, military battlefield search, natural disaster detection. In this paper,...
The number of resident space objects (RSOs) has been steadily increasing over time, posing significant risks to the safe operation on-orbit assets. accurate prediction potential collision events and implementation effective nonredundant avoidance maneuvers require precise estimation orbit positions interest propagation their associated uncertainties. Previous research mainly focuses on striking a balance between efficient computation. A recently proposed approach that integrates uncertainty...
A significant challenge in methods for anomaly detection (AD) hyperspectral images (HSIs) is determining how to construct an efficient representation anomalies and background information. Considering the high-order structures of HSIs estimation information AD, this article proposes a kernel minimum noise fraction transformation-based separation model (KMNF-BSM) separate First, spectral-domain KMNF transformation performed on original data fully mine correlation between spectral bands. Then,...
In push-broom hyperspectral imaging systems, the sensor rotation to optical plane leads linear spatial misregistration (LSM) in images (HSIs). To compensate for hardware defects through software, this paper develops four methods detect LSM HSIs. Different from traditional grayscale images, method of fitting sum abundance (FSAM) and searching equal (SEAM) are achieved by unmixing a selected rectangular transition areas containing an edge, which makes good use spectral information. The based...
The existing hyperspectral target recognition algorithms implemented on graphics processing units (GPU) have an outstanding performance in reducing the operation time. However, for push-broom imagery, real-time search not only requires high computing response and rapid decisions but also synchronizing imaging, data transmission recognition. In terms of this problem, paper proposes a new realtime method imagery. This takes advantage cross-execution concurrent execution compute unified device...
The existence of intra-class spectral variability caused by differential scene components and illumination conditions limits the improvement endmember extraction accuracy, as most algorithms directly find pixels in hyperspectral image endmembers. This paper develops a quadratic clustering-based simplex volume maximization (CSVM) approach to effectively alleviate extract CSVM first adopts spatial clustering based on simple linear iterative obtain set homogeneous partitions uses purity...