Crop Identification And Growth Monitoring Using Medium And High Resolution Remote Sensing Images

[Objective]: Based on the characteristics of the Sentinel 2 satellite image,this study investigates the species identification and growth monitoring of crops in Shihezi General Field,and discusses the application and problems of crop species identification and growth monitoring technology based on medium and high resolution remote sensing,with the aim of acquiring the crop species and growth information in Shihezi General Field and providing support for agricultural production monitoring and crop planting management in Shihezi.[Methods]: Firstly,spatial and atmospheric corrections were made to the Sentinel 2 satellite images of the Shihezi General Field to identify the major crops in the Shihezi General Field using the multi-featured object-oriented classification method of the support vector machine.To classify the growth of the main crop(cotton)at the Shihezi General Field using the natural split-point method based on satellite images of the multi-temporal Shihezi General Field and to check the accuracy of the classification.Based on the full text of the study,we discuss the application and problems of the technology for total crop class identification and growth monitoring based on medium and high resolution remote sensing images.[Results]:(1)Crops such as cotton,wheat,corn,and grapes were well separated at the Shihezi main field and had high feature integrity,with a classification accuracy of 95.4040% and a Kappa Coefficient of0.9394.Among the crops,cotton has the highest percentage of area under cultivation at 30.67%,followed by grapes,wheat and maize at 3.32%,2.68% and 2.64% respectively,while the remaining area is non-agricultural land.(2)Throughout the life cycle of cotton,its vegetation index showed a pattern of increasing and then decreasing,with a mean value of 0.102 on May 12,0.221 on June 1,0.855 on July 1,0.844 on August 10,and 0.537 on September 10.During the early growth stages of cotton,the largest proportion of cotton was poorly grown,followed by average growth.Over time,cotton growth became better,with the proportion of poorer growth decreasing and the proportion of better growth increasing from early July to mid-August,reaching a maximum in mid-August.cotton NDVI values decreased in mid-to late September.(3)Taking the medium and high satellite image as the data source,the vegetation index features and texture features related to crop species were constructed.According to the distribution trend of NDVI,the main methods of cotton growth classification were used.[Conclusion]: The main crop of Shihezi General Field is cotton,most of which has good growth potential.The multi-featured object-oriented crop identification based on medium-and high-resolution remote sensing images and the growth potential analysis based on the natural split-point method have high accuracy,which can be used for the identification and classification of crops and growth potential monitoring of small-scale remote sensing images.