Application Of Particle Swarm Optimization Algorithm In Multi-temporal Remote Sensing Image Change Detection

Remote sensing change detection is a kind of technology which quantitatively analyses the differential states and changes of target by picking up remote sensing images in different sequences in the same region.Since entering the new time and space,the theory system of remote sensing change detection has been improved and updated constantly,It can survey the resources of large area of forest,monitor and analyze the present status of land use/cover,quantitatively assess environmental disaster conditions,rationally plan urban structure,etc.It is one of the most critical technologies in the practical application of earth observation for each country,and its value is difficult to replace and beyond measure in the field of scientific application.With the development of science and technology,there is a more urgent need for the precision and automation of the remote sensing image change detection technology.Only with the high accuracy of the change detection technology can a higher level of image analysis and understanding be assisted.In this paper,the medium resolution image is used as the main experimental data,focusing on how to accurately extract the information of changing regions to a certain extent,especially on how to get the change threshold automatically,efficiently and accurately in the differential image.At present,the medium and low resolution images are mainly used for macro analysis,its state is hard to be replaced in multi-temporal remote sensing image change detection.In this paper,the medium resolution image is used as the experimental data to generate the differential image.Using the feature of intelligent optimization algorithm can find the global optimal solution in complex nonlinear multidimensional data space,and particle swarm optimization algorithm is used as the optimization tool.Then,combined with the traditional change detection method,a new change detection method is proposed.In this paper,the maximum between class variance method(Otsu)and Shannon information entropy(log entropy method,exponential entropy method,Tsallis entropy method)are improved,the main research contents include the following aspects:(1)Based on the summary and analysis of the remote sensing image change detection technology at home and abroad,change detection methods are divided into different categories,Based on each of the major categories,give a comprehensive overview on it,then the purpose and significance of the combination of particle swarm optimization algorithm and the traditional change detection method is proposed;(2)Research on improved Otsu change detection method based on particle swarm optimization algorithm:summarize the origin and basic principle of Otsu method,spread the traditional one-dimensional Otsu algorithm to multi dimension,and the spread algorithm is combined with particle swarm optimization algorithm;(3)The change detection method of Shannon information entropy based on particle swarm optimization algorithm is studied:summarize the origin and basic principle of Shannon information entropy method,spread various of information entropy algorithms to the multi threshold,then,the generalized algorithm is combined with particle swarm optimization algorithm one by one;(4)Experimental verification and accuracy evaluation.Experimental results show,change detection method based on particle swarm optimization algorithm has improved to different extents in both accuracy and time,thus,the multi temporal remote sensing image change detection method based on particle swarm optimization can obtain the global optimal threshold in the difference image generated by different time series images,and get the change information of the specified area.