Research On Remote Sensing Detection Method For Typical Ground Objects Type Change Information

The evolution of natural environment and the increase of human activities lead to the continuous changes of surface landscape structure.Getting these change information accurately in time is helpful for that environmental protection,resource management and other departments work better.The traditional method of research on the spot will be wasting time and energy,while remote sensing technology has the advantages of wide coverage,fast information acquisition,short period and large amount of information,which provides an effective technical means for remote sensing detection of change information.As an important part of remote sensing detection of change information,the research of its method has become an important application in remote sensing image analysis,which provids a theoretical reference and data support for post disaster assessment,environmental protection,agricultural production and other work.At present,the existing research methods include: color synthesis method,spectral feature variation method,difference method and ratio method and so on.With the continuous updating and development of remote sensing data and the emergence of new change detection requirements,the research about remote sensing detection method of change information is still in the stage of continuous exploration and improvement.In this paper,we selected three research areas of Greater Khingan Range Bilahe forest farm,Qinhuangdao Taolinkou reservoir and Xingcheng area,based on the multi-temporal Landsat8 OLI remote sensing images data of each research area,using the methods of multi-temporal data combination,variance analysis and factor analysis,this paper focuses on whether there are significant difference in spectral characteristic between woodland changed into burned blank,water body changed into vegetation and bare land,corn field changed into peanut field and other ground objects type in multi-temporal remote sensing images,and the remote sensing detection method of typical ground objects type change information,it can provide reference for post disaster assessment,water resources protection,agricultural planting structure adjustment and so on.The main conclusions are as follows:(1)According to the principle of multi-temporal data combination method,comparative analysis and variance analysis are used to study the difference of spectral characteristic between the typical ground objects type change information and the other ground objects type change information.The results showed that: the separation measure value of typical ground objects type and the other ground objects type in most bands is large,and the separation measure value of woodland changed into burned blank and other ground objects type in 7,6 and 12 bands reached 15833,7108 and 16923;the separation measure value of water body changed into vegetation and bare land and other ground objects type in 5,6,7 and 12 bands reached 20169,66599,16523 and 9271,the separation measure value of corn field changed into peanut field and other ground objects type in 5,12 and 14 bands reached 10529,6308 and 39693.Therefore,there are significant difference in spectral characteristic between the typical ground objects type change information and the other ground objects type change information.(2)Based on the difference of spectral characteristic between the change of typical ground objects type and other ground objects type,the "separation measure" index in the principle of variance analysis is used to quantitatively screen the optimal bands and change recognition index corresponding to the change information of different typical ground objects type.After accuracy verification,the recognition accuracy of each typical ground objects type change recognition index is 79.37%,85.04% and 75.07% respectively.The results show that the suitable recognition indexes of different typical ground objects type are different,and the multi-temporal optimal band combination method based on variance analysis can quantitatively screen out the optimal bands corresponding to different typical ground objects type change information and the suitable typical ground objects type change recognition indexes,it can effectively recognize the typical ground objects type change information.(3)Based on the principle of factor analysis and variance analysis,the typical ground objects type change recognition factors are constructed to recognize the typical ground objects type change information.From the factor load matrix,it can be found that the amount of typical ground objects type change information contained in each band is different.The weight coefficient and band can be determined according to the amount of typical ground objects type change information expressed in each band,this is the key to construct the change recognition factor of typical ground objects type;based on the principle of variance analysis,"separation measure" index can quantitatively analyze the change detection ability of each factor to the change information of typical ground objects type,and select the change recognition factor suitable for identifying the change information of typical ground objects type.After accuracy verification,the extraction accuracy of each change recognition factor is 95.35%,90.08% and90.43% respectively.Compared with the multi-temporal optimal band combination method,it makes up for the lack of quantitative indicators in the index construction link.The spectral information is more fully and reasonably used,the processing process is relatively simple,and the recognition accuracy is relatively high.The results show that the change recognition factors constructed by factor analysis method can improve the accuracy and efficiency of typical ground objects type change information recognition,and can provide reference for the quantitative research of remote sensing change detection method.