Building extraction from high resolution remote sensing images based on improved U2-Net

Huo, Zhiling; Xie, Fei; Gu, Yuchao

doi:10.1117/12.2668116

Cited by 2 publications

(2 citation statements)

References 10 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To address the problem of class imbalance in medical imaging datasets, Debesh Jha et al [11] proposed the DoubleU-Net based on the U-Net network and improved with VGG19 [12] for medical image detection in 2020, which has strong robustness and generalization ability [13] . The network structure of DoubleU-Net is shown in Figure 5.…”

Section: Basic Theory Of Doubleu-netmentioning

confidence: 99%

An Improved Method for Digital Water Meter Reading Area Segmentation Based on U~2-Net

2023

AJCIS

View full text Add to dashboard Cite

With the continuous development of intelligent technology, the recognition of water meter readings has brought convenience to people's lives. By automatically reading water meter data, work efficiency can be improved. However, due to the varying shapes and sizes of the recognition areas, it is difficult to achieve good recognition results. Therefore, in response to the problem of varying sizes of water meter reading recognition areas, this paper proposes a method for digital water meter reading recognition area segmentation based on an improved U 2 -Net. Firstly, in terms of capturing data features, an improved Double-RSU module based on the RSU module is designed. This module increases depth and complexity compared to the original RSU module, thereby improving generalization and robustness. Secondly, in terms of model training, a combination of cross-entropy loss function, Jaccard coefficient and Dice loss function is used to comprehensively evaluate the entire model by considering binary classification, segmentation accuracy and segmentation area overlap. Experimental results show thatthe proposed method has good performance in MIoU, MAE and F1Score indicators, with the MIoU indicator increasing from 0.878 to 0.91, the MAE indicator increasing from 0.001 to 0.0009, and the F1Score indicator increasing from 0.9431 to 0.9455. These results indicate that the improved model has better performance compared to the original U 2 -Net model and can more accurately perform water meter reading area segmentation.

show abstract

Section: Basic Theory Of Doubleu-netmentioning

confidence: 99%

An Improved Method for Digital Water Meter Reading Area Segmentation Based on U~2-Net

2023

AJCIS

View full text Add to dashboard Cite

show abstract

“…Therefore, this approach achieves accurate segmentation of structural plane regions in borehole images and provides insights for fully automated analysis of such images. [21][22][23][24] is an image semantic segmentation network that combines multi-scale feature extraction and a two-level nested residual network structure. It ensures the output of high-resolution feature maps while controlling memory usage and computational costs at a lower level.…”

Section: Introductionmentioning

confidence: 99%

The segmentation and intelligent recognition of structural surfaces in borehole images based on the U2-Net network

Yu,

Wang,

Cheng

et al. 2024

PLoS ONE

View full text Add to dashboard Cite

Structural planes decrease the strength and stability of rock masses, severely affecting their mechanical properties and deformation and failure characteristics. Therefore, investigation and analysis of structural planes are crucial tasks in mining rock mechanics. The drilling camera obtains image information of deep structural planes of rock masses through high-definition camera methods, providing important data sources for the analysis of deep structural planes of rock masses. This paper addresses the problems of high workload, low efficiency, high subjectivity, and poor accuracy brought about by manual processing based on current borehole image analysis and conducts an intelligent segmentation study of borehole image structural planes based on the U2-Net network. By collecting data from 20 different borehole images in different lithological regions, a dataset consisting of 1,013 borehole images with structural plane type, lithology, and color was established. Data augmentation methods such as image flipping, color jittering, blurring, and mixup were applied to expand the dataset to 12,421 images, meeting the requirements for deep network training data. Based on the PyTorch deep learning framework, the initial U2-Net network weights were set, the learning rate was set to 0.001, the training batch was 4, and the Adam optimizer adaptively adjusted the learning rate during the training process. A dedicated network model for segmenting structural planes was obtained, and the model achieved a maximum F-measure value of 0.749 when the confidence threshold was set to 0.7, with an accuracy rate of up to 0.85 within the range of recall rate greater than 0.5. Overall, the model has high accuracy for segmenting structural planes and very low mean absolute error, indicating good segmentation accuracy and certain generalization of the network. The research method in this paper can serve as a reference for the study of intelligent identification of structural planes in borehole images.

show abstract

Building extraction from high resolution remote sensing images based on improved U2-Net

Cited by 2 publications

References 10 publications

An Improved Method for Digital Water Meter Reading Area Segmentation Based on U~2-Net

An Improved Method for Digital Water Meter Reading Area Segmentation Based on U~2-Net

The segmentation and intelligent recognition of structural surfaces in borehole images based on the U2-Net network

Contact Info

Product

Resources

About