A wheat spike detection method based on Transformer

Qiong Zhou; Ziliang Huang; Shijian Zheng; Lin Jiao; Liusan Wang; Rujing Wang

doi:10.3389/fpls.2022.1023924

A wheat spike detection method based on Transformer

Front Plant Sci. 2022 Oct 20:13:1023924. doi: 10.3389/fpls.2022.1023924. eCollection 2022.

Authors

Qiong Zhou^{1

2

3}, Ziliang Huang^{1

2}, Shijian Zheng^{1

4}, Lin Jiao^{1

5}, Liusan Wang¹, Rujing Wang^{1

2}

Affiliations

¹ Institute of Intelligent Machines, Hefei Institutes of Physical Science, Chinese Academy of Sciences, Hefei, China.
² Science Island Branch, University of Science and Technology of China, Hefei, China.
³ College of Information and Computer, Anhui Agricultural University, Hefei, China.
⁴ Department of Information Engineering Southwest, University of Science and Technology, Mianyang, China.
⁵ School of Internet, Anhui University, Hefei, China.

Abstract

Wheat spike detection has important research significance for production estimation and crop field management. With the development of deep learning-based algorithms, researchers tend to solve the detection task by convolutional neural networks (CNNs). However, traditional CNNs equip with the inductive bias of locality and scale-invariance, which makes it hard to extract global and long-range dependency. In this paper, we propose a Transformer-based network named Multi-Window Swin Transformer (MW-Swin Transformer). Technically, MW-Swin Transformer introduces the ability of feature pyramid network to extract multi-scale features and inherits the characteristic of Swin Transformer that performs self-attention mechanism by window strategy. Moreover, bounding box regression is a crucial step in detection. We propose a Wheat Intersection over Union loss by incorporating the Euclidean distance, area overlapping, and aspect ratio, thereby leading to better detection accuracy. We merge the proposed network and regression loss into a popular detection architecture, fully convolutional one-stage object detection, and name the unified model WheatFormer. Finally, we construct a wheat spike detection dataset (WSD-2022) to evaluate the performance of the proposed methods. The experimental results show that the proposed network outperforms those state-of-the-art algorithms with 0.459 mAP (mean average precision) and 0.918 AP₅₀. It has been proved that our Transformer-based method is effective to handle wheat spike detection under complex field conditions.

Keywords: IoU loss function; agriculture; deep learning; transformer; wheat spike detection.