Inversion of chlorophyll content under the stress of leaf mite for jujube based on model PSO-ELM method

• DOI: 10.3389/fpls.2022.1009630
• Publication Date: 2022-09-30T12:14:51Z
• Authors: Jianqiang Lu; Hongbin Qiu; Qing Zhang; Yubin Lan; Panpan Wang; Yue Wu; Jiawei Mo; Wadi Chen; HongYu Niu; Zhiyun Wu
• Keywords: Plant culture; PSO-ELM; Plant Science; 04 agricultural and veterinary sciences; 15. Life on land; SB1-1110; SPAD; hyperspectral; damage severity; 0401 agriculture, forestry, and fisheries; jujube; SPA

During the growth season, jujube trees are susceptible to infestation by leaf mite, which reduces fruit quality and productivity. Traditional monitoring techniques for mites time-consuming, difficult, subjective, result in a time lag. In this study, method based on particle swarm optimization (PSO) algorithm extreme learning machine estimation of chlorophyll content (SPAD) under mite was proposed. Initially, image data SPAD values orchards four severities were collected analysis. Six vegetation indices value chosen correlation analysis establish model indices. To address influence colinearity between spectral bands, feature band with highest coefficient retrieved first using successive projection algorithm. modeling process, PSO initialized convergent optimal approximation fitness function value; root mean square error (RMSE) predicted measured derived as an indicator goodness-of-fit solve problems ELM weights, threshold randomness, uncertainty network parameters; finally, iterative update used determine optimize minimum or iteration number. The results reflected that significant differences observed reflectance canopy corresponding severity infestation, negatively correlated leaves. selected NDVI, RVI, PhRI, MCARI positively SPAD, whereas TCARI GI SPAD. accuracy optimized PSO-ELM (R2 = 0.856, RMSE 0.796) superior alone 0.748, 1.689). remote sensing relative shows high fault tolerance improved data-processing efficiency. provide reference utility UAV jujube.