Abstract:Spring maize is the main food crop in the alpine black soil area of Northeast China. As an effective measure to increase the final yield of crops, film mulching has the potential to be widely used in Northeast China. In order to monitor leaf area index (LAI) of spring maize under film mulching in real time and explore the influence of film mulching factors on LAI and hyperspectral reflectance of canopy, field experiments were carried out in the black soil area of Northeast China in 2019. In the experiments, three treatments, including no film mulching (M0), degradable film mulching (M1), and conventional plastic film mulching (M2) were tested. The hyperspectral reflectance data of each growth stage were obtained by hyperspectral remote sensing technology. After pre-processing the spectral data, the sensitive bands, sensitive vegetation indices and characteristic indexes of the LAI values of each film mulching treatment were extracted by correlation analysis, and the hyperspectral estimating models of LAI during the whole growth period were constructed. The results showed that the effect of film mulching on LAI was mainly before the tasseling stage. The difference in LAI between film mulching and non-mulching treatments under the same fertilization level showed a trend of first decreasing and then increasing with the growing of maize. The difference of LAI in seedling stage was the largest, and the LAI was increased by 78% under film mulching condition. The difference in hyperspectral reflectance of canopies was the largest in the middle growth period, followed by the end of growth period, and the smallest in the early growth period. Film mulching mainly affected the absorption of green and red lights by maize. The sensitive hyperspectral parameters of LAI values were different among the different film mulching treatments. Among the LAI estimation models established based on the three sensitive hyperspectral parameters (mentioned above), the models established based on the characteristic indexes had relatively high accuracy for LAI inversion results. Their fitting and verification determination coefficient R2 were above 0.80, the root mean square error RMSE were between 0.45cm2/cm2 and 0.66cm2/cm2, and the residual prediction deviation RPD were greater than 2. Film mulching had impact on LAI inversion, the inversion accuracy of LAI without film mulching was higher than that with film mulching. Model based on characteristic index was the best to monitor LAI of all treatments. Based on the characteristic index NI(722,731) showed the superiority of inversion, which was stable, accurate and effective for all film mulching treatments.
