A5056261390- Remote-Sensing Image Classification
- Remote Sensing and Land Use
- Advanced Chemical Sensor Technologies
- Machine Learning and ELM
- Face and Expression Recognition
- Human Pose and Action Recognition
- Online Learning and Analytics
- Music and Audio Processing
- Infrared Target Detection Methodologies
- Infrared Thermography in Medicine
- Thermography and Photoacoustic Techniques
- Mind wandering and attention
- Spectroscopy and Chemometric Analyses
Hubei University of Arts and Science
2014-2024
Central China Normal University
2024
Huazhong University of Science and Technology
2014
Robust and efficient infrared (IR) small target detection plays an important role in image processing for IR remote sensing. In order to detect the with high rate, low false alarm rate (FAR), speed, a novel method called high-boost-based multiscale local contrast measure (HB-MLCM) is proposed this letter. First, improved boost filter enhance frequency signal where may appear suppress signal. Then, simple MLCM further enhancing suppressing background. Finally, adaptive thresholding used...
The large number of spectral bands acquired by hyperspectral imaging sensors allows us to better distinguish many subtle objects and materials. Unlike other classical image classification methods in the multivariate analysis framework, this paper, a novel method using functional data (FDA) for accurate images has been proposed. central idea FDA is treat as continuous functions. From perspective, curve each pixel naturally viewed function. This can be beneficial making full use abundant...
Cognitive engagement involves mental and physical involvement, with observable behaviors as indicators. Automatically measuring cognitive can offer valuable insights for instructors. However, object occlusion, inter-class similarity, intra-class variance make designing an effective detection method challenging. To deal these problems, we propose the Object-Enhanced–You Only Look Once version 8 nano (OE-YOLOv8n) model. This model employs YOLOv8n framework improved Inner Minimum Point Distance...
This paper proposes a framework for hyperspectral images (HSIs) classification with composite kernels discriminant analysis (CKDA). The CKDA uses the spectral and spatial information extracted by Gaussian weighted local mean operator (GWLM) is suitable to solve few labeled samples problem of HSI, which has very important practical significance case that training are insufficient due high cost. Experimental results show GWLM can greatly improve performance, demonstrate superiority HSI in...
This article proposes a functional data discriminant analysis (FDDA) method for hyperspectral image (HSI) classification. analyzes and processes the HSI from point of view, which is novel perspective in processing. The classical methods achieve dimensionality reduction by directly eliminating redundancy data. However, proposed extracts features utilizing Functional can effectively reveal inherent characteristics with change wavelengths. Based on this, regularized weighted fitting model first...
Abstract A new spectral‐spatial hyperspectral image (HSI) classification method called hierarchical broad learning system (HBLS) has been proposed in this paper. Specifically, it combines wavelet, (BLS) and Gabor filters into a structure. First of all, wavelet is used to reduce the observation noise HSIs. Then BLS adopted acquire set pixelwise probability maps from input data, are explore spatial information by refining these maps. These two operations (BLS filtering) alternated form...
In this paper, a novel method based on wavelet transformation of functional data for accurate classification hyperspectral images is proposed. The motivation the proposed that in images, spectral curve each pixel can be viewed as function, mathematically. And it make best abundant information original images. Based this, we have an effective image which mainly used principal component analysis (FPCA) before. However, FPCA focuses global features and may not well extract local structural...