Research Article | Open Access

Volume 2023 |Article ID 0013 | https://doi.org/10.34133/plantphenomics.0013

A Novel Intelligent System for Dynamic Observation of Cotton Verticillium Wilt

Chenglong Huang,¹ Zhongfu Zhang,¹ Xiaojun Zhang,² Li Jiang,¹ Xiangdong Hua,¹ Junli Ye,² Wanneng Yang,^2,3 Peng Song ,² Longfu Zhu ^2,3

¹College of Engineering, Huazhong Agricultural University, Wuhan 430070, PR China.
²College of Plant Science & Technology, Huazhong Agricultural University, Wuhan 430070, PR China
³National Key Laboratory of Crop Genetic Improvement, National Center of Plant Gene Research (Wuhan), Huazhong Agricultural University, Wuhan 430070, PR China

Abstract

Verticillium wilt is one of the most critical cotton diseases, which is widely distributed in cotton-producing countries.  However,  the  conventional  method  of  verticillium  wilt  investigation  is  still  manual,  which  has  the disadvantages of subjectivity and low efficiency. In this research, an intelligent vision-based system was  proposed  to  dynamically  observe  cotton  verticillium  wilt  with  high  accuracy  and  high  throughput.  Firstly,  a  3-coordinate  motion  platform  was  designed  with  the  movement  range  6,100  mm ×  950  mm ×500 mm, and a specific control unit was adopted to achieve accurate movement and automatic imaging. Secondly, the verticillium wilt recognition was established based on 6 deep learning models, in which the VarifocalNet  (VFNet)  model  had  the  best  performance  with  a  mean  average  precision  (mAP)  of  0.932.  Meanwhile,  deformable  convolution,  deformable  region  of  interest  pooling,  and  soft  non-maximum  suppression optimization methods were adopted to improve VFNet, and the mAP of the VFNet-Improved model improved by 1.8%. The precision–recall curves showed that VFNet-Improved was superior to VFNet for each category and had a better improvement effect on the ill leaf category than fine leaf. The regression results  showed  that  the  system  measurement  based  on  VFNet-Improved  achieved  high  consistency  with  manual  measurements.  Finally,  the  user  software  was  designed  based  on  VFNet-Improved,  and  the  dynamic  observation  results  proved  that  this  system  was  able  to  accurately  investigate  cotton  verticillium wilt and quantify the prevalence rate of different resistant varieties. In conclusion, this study has demonstrated a novel intelligent system for the dynamic observation of cotton verticillium wilt on the seedbed, which provides a feasible and effective tool for cotton breeding and disease resistance research.