A Novel Multi-Scale Attention PFE-UNet for Forest Image Segmentation

Boyang Zhang,Mingyu Gao,Jianfeng Chen,Haiming Ni,Hong Yang,Hongbo Mu,Dawei Qi

doi:10.3390/f12070937

Abstract

The precise segmentation of forest areas is essential for monitoring tasks related to forest exploration, extraction, and statistics. However, the effective and accurate segmentation of forest images will be affected by factors such as blurring and discontinuity of forest boundaries. Therefore, a Pyramid Feature Extraction-UNet network (PFE-UNet) based on traditional UNet is proposed to be applied to end-to-end forest image segmentation. Among them, the Pyramid Feature Extraction module (PFE) is introduced in the network transition layer, which obtains multi-scale forest image information through different receptive fields. The spatial attention module (SA) and the channel-wise attention module (CA) are applied to low-level feature maps and PFE feature maps, respectively, to highlight specific segmentation task features while fusing context information and suppressing irrelevant regions. The standard convolution block is replaced by a novel depthwise separable convolutional unit (DSC Unit), which not only reduces the computational cost but also prevents overfitting. This paper presents an extensive evaluation with the DeepGlobe dataset and a comparative analysis with several state-of-the-art networks. The experimental results show that the PFE-UNet network obtains an accuracy of 94.23% in handling the real-time forest image segmentation, which is significantly higher than other advanced networks. This means that the proposed PFE-UNet also provides a valuable reference for the precise segmentation of forest images.

Highlights

IntroductionForest resources are highly relevant to national economic development [1], and play an essential role in many aspects, such as ecotourism, landscape construction, and maintenance of ecological security of the country [2,3,4,5]
The Pyramid Feature Extraction module (PFE)-UNet network is composed of two parts, forming a symmetrical structure: the network is composed of two parts, forming a symmetrical structure: encoder and the decoder part
To evaluate the performance of the proposed model, the networks segmentation results are compared with the corresponding ground truth and the comparison results for each pixel are grouped into true positive (TP), false positive (FP), false negative (FN), and true negative

Summary

Introduction

Forest resources are highly relevant to national economic development [1], and play an essential role in many aspects, such as ecotourism, landscape construction, and maintenance of ecological security of the country [2,3,4,5]. The impact of overexploitation and other reasons has led to a serious shortage of forest resources in China, and the ecological environment has become extremely fragile [6,7]. Traditional image segmentation methods suffer from low accuracy and time consumption, which render it hard to handle segmentation tasks in complex scenes. How to handle the segmentation task in complex scenes has been one of the most challenging problems within the field of forest image segmentation

Methods

Results

Discussion

Conclusion