Farmland boundary extraction based on the AttMobile-DeeplabV3+ network and least squares fitting of straight lines

Hao Lu; Hao Wang; Zhifeng Ma; Yaxin Ren; Weiqiang Fu; Yongchao Shan; Shupeng Hu; Guangqiang Zhang; Zhijun Meng

doi:10.3389/fpls.2023.1228590

Farmland boundary extraction based on the AttMobile-DeeplabV3+ network and least squares fitting of straight lines

Front Plant Sci. 2023 Aug 18:14:1228590. doi: 10.3389/fpls.2023.1228590. eCollection 2023.

Authors

Hao Lu^{1

2

3}, Hao Wang^{1

2

3}, Zhifeng Ma⁴, Yaxin Ren⁵, Weiqiang Fu^{1

2

3}, Yongchao Shan^{1

2

3}, Shupeng Hu^{1

2

3}, Guangqiang Zhang^{1

2

3}, Zhijun Meng^{1

2

3}

Affiliations

¹ Intelligent Equipment Research Center, Beijing Academy of Agriculture and Forestry Sciences, Beijing, China.
² State Key Laboratory of Intelligent Agricultural Power Equipment, Beijing, China.
³ National Engineering Research Center of Intelligent Equipment for Agriculture, Beijing, China.
⁴ School of Integrated Circults and Electonics, Beijing Institute of Technology, Beijing, China.
⁵ Beijing Research Center for Information Technology in Agriculture, Beijing Academy of Agriculture and Forestry Sciences, Beijing, China.

Abstract

The rapid extraction of farmland boundaries is key to implementing autonomous operation of agricultural machinery. This study addresses the issue of incomplete farmland boundary segmentation in existing methods, proposing a method for obtaining farmland boundaries based on unmanned aerial vehicle (UAV) remote sensing images. The method is divided into two steps: boundary image acquisition and boundary line fitting. To acquire the boundary image, an improved semantic segmentation network, AttMobile-DeeplabV3+, is designed. Subsequently, a boundary tracing function is used to track the boundaries of the binary image. Lastly, the least squares method is used to obtain the fitted boundary line. The paper validates the method through experiments on both crop-covered and non-crop-covered farmland. Experimental results show that on crop-covered and non-crop-covered farmland, the network's intersection over union (IoU) is 93.25% and 93.14%, respectively; the pixel accuracy (PA) for crop-covered farmland is 96.62%. The average vertical error and average angular error of the extracted boundary line are 0.039 and 1.473°, respectively. This research provides substantial and accurate data support, offering technical assistance for the positioning and path planning of autonomous agricultural machinery.

Keywords: DeeplabV3+; UAV remote sensing; farmland boundary extraction; linear fitting; semantic segmentation.

Grants and funding

This work was supported by the National Key Research and Development Program of China (Grant No. 2020YFB1709600), the Youth Fund of Beijing Academy of Agricultural and Forestry Sciences (Grant No. QNJJ202320 &QNJJ202103 ), National Key Research and Development Program of China (Grant No. 2021YFD2000600), Key Research and Development Program of Shandong Province (Grant No. 2022SFGC0202), Chen Liping Expert Workstation of Yunnan Province (Grant No. 202105AF150030), and Outstanding Young Talents Projects of Beijing Academy of Agriculture and Forestry Sciences.