An Automatic Change Detection Technology for remote sensing data using Gaussian Mixture Model

It is a hotspot in the field of remote sensing image analysis and application by using the macro and realtime features of the remote sensing image data for its change in Land and Resources. This paper introduces an automatic change monitoring method with remote sensing image data and historical interpretation vector data, which is based on the Gaussian Mixture Model and the vector-guided image spot segmentation technique. With the remote sensing image data of GF-1 in November 2014 and the historical remote sensing interpretation data of the corresponding region in 2014, we test our model in the image coverage of 400 square Km, which is located in Yalujiang Reserve in Dandong City, Liaoning Province. For the class of dry land, water body and vegetation surface, the monitoring rate was over 90% and the missed rate was less than 10%. Experiments show that the proposed method can obviously improve precision and efficiency and meet the production needs.