| Dynamic monitoring of crop growth can keep abreast of the crop growth status, seedlings, soil moisture, nutriture and its changes. Then, a variety of management strategies will be taken to ensure the crop normal growth. At the same time, the monitoring can also provide the losses when the meteorological disasters, pests or diseases happen. In all, the crop growth dynamic monitoring is of great importancee.In this paper, systematic study has been done to monitor the growth of Beijing winter wheat. The data sources are three HSI (Hyperspectral Image) of environmental satellites gotten in 2009, multispectral images and ground hyperspectral data and physical and chemical parameters gotten in 2010. The main works are as follows:(1)Data acquisition and processing. This includes the acquisition and processing of ground hyperspectral data, HSI hyperspectral images, multispectral images, physical and chemical parameters such as LAI (Leaf Area Index), chlorophyll, soluble sugar, leaf total nitrogen and so on. The Remote sensing images processing includes geometric correction, radiometric calibration, and atmospheric correction.(2)The Gaussian function is taken to simulate the response function of HSI, and the ground hyperspectral data is matched to the HSI channels. The correlative analysis between the original spectra, derivative spectra, normalized spectral index (NDSI), subtraction Index (SSI) of the ground ASD data and its simulating HSI and LAI, chlorophyll a, chlorophyll b, chlorophyll a + b, soluble sugar and total nitrogen are taken. After choosing the sensitive bands, the statistical model is established.(3)The red edge parameters were extracted from similating HSI, and the correlation analysis between the red edge parameters and physical and chemical parameters of winter wheat were carried on. The models based on red edge width, NDSI to predict leaf nitrogen content and LAI respectively, and based on SSI to predict chlorophyll a , chlorophyll b, chlorophyll a + b and soluble sugars were constructed. Further, carbon and nitrogen ratio were calculated by soluble sugars and nitrogen. The mapping of the various physical and chemical parameters of experimental area was done by the HSI image. The growth monitoring is done using LAI, chlorophyll a, chlorophyll b, chlorophyll a + b, carbon and nitrogen ratio. At last, the RGB false color was composed with LAI, chlorophyll, carbon and nitrogen ratio.(4)Eight vegetation indexes are constructed by matched spectral reflectance, and the correlative analysis was done to choose the best index. The statistical model was established between the selected index and LAI. Three growth stages LAI were inversed and verified. Then the growth monitoring was done for the classification and grading of LAI.(5)NIR-R spectral was constructed by CCD data of environment satellite to calculate the slope of the soil and ,the vertical dry vegetation index (PDI) . Based on the surface emissivityεand atmospheric moisture contentω, the canopy temperature of winter wheat and the temperature of vegetation drought index (TVDI) were calculated by the thermal infrared channel method. The classification and grading was done on the two kinds of drought indexes to monitor the drought, also comparison were done between them.Monitoring crop growth using HJ-1 HSI data, the accuracy are as follows: LAI is the highest with verification accuracy is 0.8601, carbon to nitrogen ratio is second with 0.8008 accuracy, the lowest is chlorophyll, the accuracy of chlorophyll b, chlorophyll a + b, chlorophyll a are 0.6224 , 0.533,0.4544 respectively. Three growth stages LAI are predicted using CCD data, the verified accuracy are 0.8145,0.6789,0.8502. Based on the soil moisture, the inversed spatial distribution and levels of temperature Vegetation Dryness Index (TVDI) and perpendicular drought index ( PDI) are consistent with the actual situation, and the monitoring results are consistent. In all, those show that monitoring growth and drought are feasible using the HSI, CCD, and IRS data of HJ-1... |
PDF Full Text Request