A5061758518- Remote Sensing in Agriculture
- Remote-Sensing Image Classification
- Wood and Agarwood Research
- Fuel Cells and Related Materials
- Advanced Image Fusion Techniques
- Electrocatalysts for Energy Conversion
- Neural Networks and Applications
- Autonomous Vehicle Technology and Safety
- Machine Learning and ELM
- Advancements in Solid Oxide Fuel Cells
- Advanced Neural Network Applications
Qiqihar University
2021-2023
First Automotive Works (China)
2023
Northeast Forestry University
2021-2022
Activation functions are crucial in deep learning networks, given that the nonlinear ability of activation endows neural networks with real artificial intelligence. Nonlinear nonmonotonic functions, such as rectified linear units, Tan hyperbolic (tanh), Sigmoid, Swish, Mish, and Logish, perform well models; however, only a few them widely used mostly all applications due to their existing inconsistencies. Inspired by MB-C-BSIF method, this study proposes Smish, novel function, expressed...
As a current research hotspot, graph convolution networks (GCNs) have provided new opportunities for tree species classification in multi-source remote sensing images. To solve the challenge of limited label information, model was proposed by using semi-supervised fusion method hyperspectral images (HSIs) and multispectral (MSIs). In model, graph-based attribute features pixel-based are fused to deepen correlation improve accuracy. Firstly, employs canonical analysis (CCA) maximize images,...
In order to solve the problem of manual labeling in semi-supervised tree species classification, this paper proposes a pixel-level self-supervised learning model named M-SSL (multisource learning), which takes advantage information plenty multisource remote sensing images and methods. Based on hyperspectral (HSI) multispectral (MSI), features were extracted by combining generative methods with contrastive Two kinds encoders MAAE AAE encoder) MVAE VAE proposed, respectively, set up pretext...
Multi-source data remote sensing provides innovative technical support for tree species recognition. Tree recognition is relatively poor despite noteworthy advancements in image fusion methods because the features from multi-source each pixel same region cannot be deeply exploited. In present paper, a novel deep learning approach hyperspectral imagery proposed to improve accuracy classification of species. The method, named double branch (DBMF) could more determine relationship between and...
ALS (Automatic Headlamp Leveling System) is an intelligent headlamp control system that can adjust the distance of low beam illumination in real time, providing safer driving vision for vehicle drivers and passengers dark environments, improving comfort safety. This paper proposes a machine learning method(Decision Tree) to assist deciding whether based on current road type by reading various sensor data vehicle. solution attempt implement using methods, advanced technology make cars smarter safer.