Hyperspectral imaging combined with CNN for maize variety identification

Fu Zhang; Fangyuan Zhang; Shunqing Wang; Lantao Li; Qiang Lv; Sanling Fu; Xinyue Wang; Qingfeng Lv; Yakun Zhang

doi:10.3389/fpls.2023.1254548

Hyperspectral imaging combined with CNN for maize variety identification

Front Plant Sci. 2023 Sep 8:14:1254548. doi: 10.3389/fpls.2023.1254548. eCollection 2023.

Authors

Fu Zhang^{1

2}, Fangyuan Zhang¹, Shunqing Wang¹, Lantao Li³, Qiang Lv⁴, Sanling Fu⁵, Xinyue Wang¹, Qingfeng Lv¹, Yakun Zhang¹

Affiliations

¹ College of Agricultural Equipment Engineering, Henan University of Science and Technology, Luoyang, China.
² Collaborative Innovation Center of Machinery Equipment Advanced Manufacturing of Henan Province, Henan University of Science and Technology, Luoyang, China.
³ College of Resources and Environment, Henan Agricultural University, Zhengzhou, China.
⁴ College of Agriculture / Peon, Henan University of Science and Technology, Luoyang, China.
⁵ College of Physical Engineering, Henan University of Science and Technology, Luoyang, China.

Abstract

Introduction: As the third largest food crop in the world, maize has wide varieties with similar appearances, which makes identification difficult. To solve the problem of identification of hybrid maize varieties, a method based on hyperspectral image technology combined with a convolutional neural network (CNN) is proposed to identify maize varieties.

Methods: In this study, 735 maize seeds from seven half-parent hybrid maize varieties were regarded as the research object. The maize seed images in the range of 900 ~ 1700nm were obtained by hyperspectral image acquisition system. The region of interest (ROI) of the embryo surface was selected, and the spectral reflectance of maize seeds was extracted. After Savitzky-Golay (SG) Smoothing pretreatment, Maximum Normalization (MN) pretreatment was performed. The 56 feature wavelengths were selected by Competitive Adaptive Reweighting Algorithm (CARS) and Successive Projection Algorithm (SPA). And the 56 wavelengths were mapped to high-dimensional space by high-dimensional feature mapping and then reconstructed into three-dimensional image features. A five-layer convolution neural network was used to identify three-dimensional image features, and nine (SG+MN)-(CARS+SPA)-CNN maize variety identification models were established by changing the input feature dimension and the depth factor size of the model layer.

Results and discussion: The results show that the maize variety classification model works best, when the input feature dimension is 768 and the layer depth factor d is 1.0. At this point, the model accuracy of the test set is 96.65% and the detection frame rate is1000 Fps/s in GPU environment, which can realize the rapid and effective non-destructive detection of maize varieties. This study provides a new idea for the rapid and accurate identification of maize seeds and seeds of other crops.

Keywords: convolution neural network; high dimensional feature mapping; hyperspectral imaging technology; maize; non-destructive testing.

Grants and funding

This Paper is supported by the National Natural Science Foundation of China (Grant No. 52075149), Frontier Exploration Projects of Longmen Laboratory (Grant No. LMQYTSKT032), the National Key Research and Development Program of China project (No. 2017YFD0301106), the Scientific and Technological Project of Henan Province (Grant No. 232102111119 and 222102110196), Colleges and Universities of Henan Province Youth Backbone Teacher Training Program (Grant No. 2017GGJS062).