Soybean yield estimation based on hyperspectral technology under water stress

Jindazaohuang No. 2 was selected as the test material to analyze the law of yield change under different water stress conditions. Through spectral monitoring of soybean canopy height under different water stress treatments, the rule of yield variation was explored and the response rule of canopy spectral reflectance to water stress was studied. At the same time, successive projections algorithm is used to extract the characteristic band, and the multiple linear regression monitoring model based on the soybean yield characteristic band is constructed. The results showed that the higher the degree of water stress, the lower the yield of soybean is basically. The canopy spectral reflectance is significantly related to yield of soybean, the coefficient of determination R2 of the successive projections algorithm-multiple linear regression model 41 days after construction was 0.791. The model test results are significant and the fitting effect is good. The research found that it is feasible to use hyperspectral technology to monitor the growth and yield of soybean under water stress quickly and nondestructively.

Automated summary

This study analyzed the law of yield change under different water stress conditions using Jindazaohuang No. 2 as the test material. By monitoring the soybean canopy height under different water stress treatments, the rule of yield variation was explored and the response rule of canopy spectral reflectance to water stress was studied. The study found that the higher the degree of water stress, the lower the yield of soybean. The research also constructed a multiple linear regression monitoring model based on soybean yield characteristic band, with a coefficient of determination R2 of 0.791.