Research on Area Information Extraction of Cotton Field Based on UAV Visible Light Remote Sensing

【Objective】 This project aims to use object-oriented image classification method to extract the planting information of the visible light remote sensing image of the UAV in the hope of providing a new method for extracting large-scale farmland information and improving the speed and precision of classification results.【Method】The study selected fixed-wing UAV equipped with a camera and obtained the visible light images of 135th regiment farm of the eighth division of Xinjiang Production and Construction Corps. With the help of eCognition software platform, using the object-oriented method, the cotton planting information in the study area was extracted for experiments.【Result】The planting area of cotton extracted by visual interpretation was 0.35 km²,and that by object-oriented approach was 0.33 km². The results showed that this method could effectively extract the cotton planting area in the study area, and the classification accuracy reached 94.29%, error of 5.71%.【Conclusion】Compared with traditional pixel-based classification methods, using the object-oriented classification method to extract the range information of visible light images captured by UAV has higher extraction accuracy and is greatly closer to visual interpretation.