Analysis Of Landscape Pattern And Information Extraction Of Reclamation Area In Jinzhou Bay Based On Object-oriented Technology

Based on the data of high resolution two satellite remote sensing images,the sea area near Jinzhou Bay,including Jinzhou and Huludao coast,was intercepted by the ENVI remote sensing image clipping function.Based on the object oriented software eCognition Developer 8.7 under the support of multi-scale segmentation,classification and feature selection,establish the corresponding fuzzy classification and nearest neighbor classification rules and reclamation activities,types of information extraction based on object oriented operation of the study area,the study area get high classification accuracy and the reclamation activity classification information;at the same time using the ENVI remote sensing data processing software to cover the entire area of the 2014 Landsat 8 satellite remote sensing image reclamation type pixel based supervised classification,and the classification results of pattern index of reclamation activity type classification map analysis of landscape pattern feature extraction,according to the study area reclamation activities landscape pattern status quo Accordingly,it is suggested to provide technical support for the planning of reclamation activities and the rapid and accurate acquisition of remote sensing classification information.The following conclusions are drawn from this study:(1)the object-oriented classification method is feasible for extracting the information of reclamation activities,and it can improve the accuracy of information extraction.In this paper,based on the high score two image data in 2015 and the Landsat8 image data in the year of 2014,the object-oriented and pixel based classification of the reclamation information.The results show that the classification results of object oriented analysis method the overall accuracy is 93.52%,Kappa coefficient is 0.92,compared to the classification results based on pixels improves the overall accuracy by 0.26%,Kappa coefficient is increased by 0.28%,improve the interpretation accuracy.The object oriented classification method can acquire the information of the study area more accurately.The reclamation of the extraction results of various types of information shows that in 2015 the scope of the study the purpose of confining coastal industry accounted for the largest reclamation type,an area of 138.81 hectares,accounting for 42.44% of the area of the reclamation area,followed by filling without construction land,an area of 72.95 hectares,the area accounted for 22.30% of the total area of land reclamation the use of sea salt,an area of 27.74 hectares,is located in the third,the study area accounted for a total area of 8.48%.(2)In this paper,landscape pattern index and landscape level index were used to analyze the landscape pattern index of reclamation area in 2010 and 2015.The results show that: landscape level,2010 to 2015 years in waters near the Jinzhou Bay reclamation activities change the number of landscape patches is obvious,the biggest change is the industrial land types,the number of patches from 2010 10 increased to 74 in 2015,followed by filling without construction land types,land use types is again urban construction;landscape type accounted for the largest percentage of reclamation type is the coastal industry,increased from 32.84% in 2010 to 40.74% in 2015,an average increase of 0.24 percentage points,followed by the fishery and fill in the sea without building types;the landscape level,the Shannon diversity index(SHDI)increased from 1.74 in 2010 to in 2015 1.83;Shannon evenness index(SHEI)decreased slightly from 0.75 in 2010 to 0.74 in 2015,the landscape spread upward trend larger and juxtaposition index increased from 49.83% in 2010.As high as 69.02% in 2015,the reclamation index(CONTAG)of the study area decreased slightly from 2010 to 55.33% in the year of 2015 to 51.31%.