| At present,China’s agricultural technology continues to progress,smart agriculture and precision agriculture have become the main trend of future agricultural development,in recent years,with the continuous update and development of optical imaging systems and image processing technology,the phenotypic acquisition method of analyzing crop growth and development from the image level has gradually been recognized by breeders.As an important food crop,wheat is one of the most widely distributed food crops in the world,occupying a very important position in China’s agricultural production,how to break the bottleneck of traditional phenotypic acquisition,achieve high-throughput,non-destructive and accurate dynamic analysis through high-throughput crop phenotyping platform,and further explore its potential relationship with genetic mechanism,which is of great significance in wheat genetic breeding research.In this paper,the images of the whole growth period of wheat were acquired by the lidar and the dual-angle RGB camera equipped with the high-throughput phenotyping platform in the field,and the phenotypic acquisition and analysis strategy of the whole growth period of wheat based on the image level was proposed,and the high-throughput phenotypic information obtained was further analyzed to explore the association between phenotype and genotype.The main research work is as follows:(1)Li DAR was used to obtain point cloud images for noise reduction,the top height of the canopy was calculated by the median method and the cumulative distribution method was used to calculate the ground soil height,and the difference was calculated to obtain high-throughput plant height phenotypic data of wheat at different growth stages.The reliability of the plant height phenotypic traits was evaluated by the control cultivars,and the R~2was above 0.97.It effectively confirmed the reliability and accuracy of lidar measurement in wheat crops.(2)Dynamic genome-wide association analysis of high-throughput plant height phenotypic data obtained by lidar at different growth stages,a total of 72 significant SNP sites were detected on 15 chromosomes in three periods,and the three dwarf stem genes that controlled plant height were Rth-B1b(Rth1)on chromosome 4B,Rth-D1b(Rth2)on chromosome 4D and Rth12 on chromosome 5A.From the side,the feasibility of high-throughput measurement of wheat phenotypic data by lidar was proved.(3)Using the two-angle visible light image obtained by RGB,GF(green fraction)was extracted by EXG model and Output threshold algorithm,and the GAI(Green area index)value of different RGB images under 0°and 45°dual viewing angles was estimated according to the turbid medium model and neural network inverted GF,and the reliability evaluation R~2of the estimated value and the measured value was 0.87,and the estimation effect was good.(4)Based on the law of allometric growth,a power function biomass model for estimating the vegetative growth stage of wheat by GAI was established,and the reliability of the model was verified,and the R~2was 0.76 and the RMSE was0.62ton/ha,which was of significant significance.In conclusion,a genotypic acquisition and analysis strategy of wheat whole growth period based on field phenotyping platform is proposed,and the feasibility of this strategy is verified by genome-wide association and biomass prediction models.This paper provides a new technology and thinking direction for phenotypic extraction of crops in the whole growth period. |
PDF Full Text Request