| Inefficient water and fertilizer utilization is an important factor limiting agricultural production in the river-loop irrigation area,while controlled-release fertilizer makes nitrogen utilization efficiency significantly higher due to its slow-release nitrogen characteristics.Traditional farmland nitrogen deficit monitoring is time-consuming and laborious,while UAVs have the advantages of rapid,high-precision and automated monitoring,and play an important role in realizing smart fertilization of farmland.In this paper,the experimental study of different controlled-release fertilizer dosage on summer maize growth was conducted in the central part of the river-loop irrigation area from2021 to 2022 with five treatments of no fertilizer(0,N0),conventional fertilizer(225,N1),and controlled-release fertilizer with low(150,N2),medium(225,N3)and high(300,N4)kg/hm~2.The effects of different fertilizers and different controlled-release fertilizer dosages on the growth parameters of summer maize at different fertility stages were analyzed,and the correlation between the leaf area index(LAI),leaf nitrogen content(LNC)and chlorophyll content(SPAD)of summer maize canopy and the spectral variables of UAV remote sensing were revealed,and a multispectral remote sensing model of LAI,LNC and SPAD was constructed for the controlled-release fertilizer applied farmland,as well as the above-ground biomass,LAI,SPAD using different fertility stages.The model of summer maize yield estimation using above-ground biomass,LAI,SPAD and vegetation index at different fertility stages was established to clarify the response characteristics of canopy LAI,LNC and SPAD distribution to nitrogen fertilizer type and controlled-release fertilizer dosage at the nodulation,tasseling and filling stages of summer maize,and optimize the on-farm nitrogen application program according to the influence mechanism of different controlled-release fertilizer amounts on yield,which provides important support for accurate and intelligent fertilization of summer maize.The main findings are as follows:(1)There were no significant difference in the growth parameters of summer maize in the early stage of growth,but in the later stage,nitrogen was gradually released with the controlled release fertilizer,and there was a significant difference between the treatments(P<0.05).In the late growth stage of summer maize,under the same nitrogen application treatment,the plant height of controlled release fertilizer treatment was11.45%higher than that of traditional fertilizer treatment in two years,the stem diameter was increased by 19.54%,the aboveground biomass was increased by 20.36%,and the dry matter weight was increased by 9.29%,LAI was increased by 12.78%,LNC was increased by 16.74%,SPAD was increased by 14.19%.This shows that the application of controlled-release fertilizers is more conducive to improving the growth index of summer maize and promoting crop growth and development than traditional fertilizers.But the amount of controlled release fertilizer should not exceed 225kg/hm~2.(2)The correlation between maize canopy vegetation index and LAI,LNC and SPAD at different growth stages could reach significant correlation(P<0.01).The correlation between vegetation indices MSR,NDVI,GNDVI,RVI,OSAVI and MNLI and maize LAI,LNC and SPAD was good,which could reach more than 0.6.The above vegetation indices could be well used to construct LAI,LNC and SPAD estimation models.(3)The optimal number of spectral variables of LAI,LNC and SPAD stepwise regression models based on UAV multi-spectral images in each period of 2021 and 2022ranged from 4 to 8,and the average R~2of modeling set and validation set of LAI,LNC and SPAD estimation models in each growth period was greater than 0.6.(4)Based on aboveground biomass,LAI,SPAD and vegetation index,the optimal yield estimation model of summer maize UAV remote sensing was constructed.The model based on vegetation index had the highest accuracy,which was 8.05%,11.56%and 7.03%higher than that based on aboveground biomass,LAI and SPAD,respectively.The determination coefficient of the correlation model between vegetation index and yield in the whole growth period was significantly higher than that of the model established in a single growth period.The order of determination coefficient was heading stage,filling stage and jointing stage.In addition,the optimal vegetation index estimation model for the whole growth period is the NDVI estimation model,and the two-year model R~2is 0.72 and 0.79,respectively.(5)Under the same nitrogen application rate,the yield of maize in the plot with controlled release fertilizer from 2021 to 2022 was 7.97%and 10.21%higher than that of traditional fertilizer,respectively.Using the quadratic regression analysis model,the optimal controlled-release fertilizer nitrogen application rates for 2021 and 2022 yields were 208.78kg/hm~2and 201.04kg/hm~2,respectively.It can be estimated that the maize yields under the optimal nitrogen application rates were 11585.37 kg/hm~2and 11907.73kg/hm~2,respectively. |
PDF Full Text Request