Abstract

Change detection (CD) is essential to the accurate understanding of land surface changes using available Earth observation data. Due to the great advantages in deep feature representation and nonlinear problem modeling, deep learning is becoming increasingly popular to solve CD tasks in remote-sensing community. However, most existing deep learning-based CD methods are implemented by either generating difference images using deep features or learning change relations between pixel patches, which leads to error accumulation problems since many intermediate processing steps are needed to obtain final change maps. To address the above-mentioned issues, a novel end-to-end CD method is proposed based on an effective encoder-decoder architecture for semantic segmentation named UNet++, where change maps could be learned from scratch using available annotated datasets. Firstly, co-registered image pairs are concatenated as an input for the improved UNet++ network, where both global and fine-grained information can be utilized to generate feature maps with high spatial accuracy. Then, the fusion strategy of multiple side outputs is adopted to combine change maps from different semantic levels, thereby generating a final change map with high accuracy. The effectiveness and reliability of our proposed CD method are verified on very-high-resolution (VHR) satellite image datasets. Extensive experimental results have shown that our proposed approach outperforms the other state-of-the-art CD methods.

Highlights

  • With ever-increasing Earth observation data available from all kinds of satellite sensors, such as DeepGlobal, WorldView, QuickBird, ZY-3, GF1, GF2, Sentinel, and Landsat, it is easy to obtain multi-temporal remote sensing (RS) data using the same or different sensors

  • To verify the superiority and effectiveness of our proposed change detection (CD) method, some SOTA image-based deep learning change detection (IB-DLCD) approaches are compared, which are described as follows: 1) Change detection network (CDNet) [56] was proposed for pixel-wise CD in street view scenes, which consists of contraction blocks and expansion blocks, and the final change map (CM) is generated by a soft-max layer

  • To verify the superiority and effectiveness of our proposed CD method, some SOTA IB-DLCD approaches are compared, which are described as follows: (1) Change detection network (CDNet) [56] was proposed for pixel-wise CD in street view scenes, which consists of contraction blocks and expansion blocks, and the final CM is generated by a soft-max layer

Read more

Summary

Introduction

With ever-increasing Earth observation data available from all kinds of satellite sensors, such as DeepGlobal, WorldView, QuickBird, ZY-3, GF1, GF2, Sentinel, and Landsat, it is easy to obtain multi-temporal remote sensing (RS) data using the same or different sensors. Based on the analysis unit, traditional CD methods can be divided into two categories: pixel-based CD (PBCD) and object-based CD (OBCD) [6]. In the former case, a difference image (DI) is usually generated by directly comparing pixel spectral or textual values, from which the final change map (CM) is obtained by threshold segmentation or cluster analysis. In order to make finer predictions, some methods, such as atrous convolution [61], residual connections [62], and pyramid pooling modules [63], utilize intermediate layers to enhance the output feature maps, which contribute to expanding the receptive field and overcoming the vanishing gradient problems

Methods
Results
Conclusion
Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call