Abstract:High-resolution remote sensing images have become mainstream remote sensing data, but there is an obvious "salt and pepper phenomenon" in the existing semantic segmentation methods of high-resolution remote sensing images. The purpose of this paper is to propose an improved deep convolutional neural network based on HRNet and PSPNet to segment and realize deep scene analysis and improve the pixel-level semantic segmentation representation of high-resolution remote sensing images. Based on hierarchical multiscale segmentation technology research, the main method is multiband segmentation; the vegetation, buildings, roads, waters and bare land rule sets in the experimental area are established, the classification is extracted, and the category is labeled at each pixel in the image. Using the image classification network structure, different levels of feature vectors can be used to meet the judgment requirements. The HRNet and PSPNet algorithms are used to analyze the scene and obtain the category labels of all pixels in an image. Experiments have shown that artificial intelligence uses the pyramid pooling module in the classification and recognition of CCF satellite images. In the context of integrating different regions, PSPNet affects the region segmentation accuracy. FCN, DeepLab and PSPNet are now the best methods and achieve 98% accuracy. However, the PSPNet object recognition algorithm has better advantages in specific areas. Experiments show that this method has high segmentation accuracy and good generalization ability and can be used in practical engineering.

Research on BP Neural Network for Terahertz Image Segmentation

Improved Pulse Coupled Neural Network for Target Segmentation in Infrared Images

A Terahertz Image Super-Resolution Reconstruction Algorithm Based on the Deep Convolutional Neural Network

All-optical Diffractive Neural Networked Terahertz Hologram.

Image Segmentation Based on Optimized BP Neural Network

An Image Segmentation Approach Based on Fuzzy-Neural-network Hybrid System

Terahertz Image Super-Resolution based on a Complex Network with Wavelet and Attention Mechanisms

Identification of blurred terahertz images by improved cross-layer convolutional neural network

Hybrid Multi-Head Physics-informed Neural Network for Depth Estimation in Terahertz Imaging

Fuzzy neighbourhood neural network for high-resolution remote sensing image segmentation

Efficient sub-pixel convolutional neural network for terahertz image super-resolution

A VQ-VAE Based Recognition Algorithm for Terahertz Security Inspection Images

Supervised Image Segmentation for High Dynamic Range Imaging

Grayscale Image Segmentation With Quantum-Inspired Multilayer Self-Organizing Neural Network Architecture Endorsed by Context Sensitive Thresholding

Texture Image Segmentation Method Based on Fractal Theory and Kohonen Neural Network

Optimal Imaging and Semantic Segmentation of the Human Body for Passive THz Security Inspection

Method of Computer-Generated Hologram Compression and Transmission Using Quantum Back-Propagation Neural Network

Image Compression of BP Neural Network with Hyperbolic Tangent Non-linear Function and Adding Mean Quantization Algorithm

HRNet- and PSPNet-based multiband semantic segmentation of remote sensing images

Image Compression of BP Neural Network with SPIHT Algorithm

THz time-domain spectral imaging enhancement based on an MLP-SRCNN composite network