Automatic Tea Garden Identification Method Based on Cloud Platform Fusion of Multi-Source Satellite Images and Tea Phenological Period

The invention belongs to the technical field of remote sensing target recognition, and discloses an automatic tea garden recognition method based on cloud platform fusion of multi source satellite images and tea tree phenology. The method combines all Landsat 7/8 and Sentinel-2A/B satellite images in a study area in the study period, and automatically recognizes tea gardens by extracting phenological differences between tea and other land types in different periods; firstly, an evergreen vegetation mask is generated; secondly, the growth period of tea is divided into seven periods, and the phenological index for classification is extracted according to the high-quality time series curve; finally, the tea gardens in the study area are extracted pixel by pixel. The method makes full use of the unique phenological index caused by the artificial management mode and artificial planting mode in the tea garden, and is more in line with the real growth law of the tea garden; all satellite images in the study area and study period are integrated, which is beneficial to capture the key phenological period of tea garden and effectively improves the accuracy of tea garden identification. Figures Landsat 7/8 Senlinel-2 A/B Surface reflection data Surface reflection data Eliminoionofinferior observation imagecoordnation Vegetation index (NDV, LSWI) MCTA, PCTA, OEVA NDVI time sequence Generate EVAmask in 2019 Savitzky-Golay Sythlesisea10 days Linear interpolation Filter-smoothing Recognize MCTA. PCTA. DEVA Figure 1 11341'E ii35E 114°'t 114 i' 114°20' a) de) aM EVA 1F ukre Figure 2 If3-40VE 113-SOE 14?0T I14-IOT 11ie2O'I a) d) 0 to 24km Figure 3