Research On Road Information Extraction From High Resolution Remote Sensing Images

With the rapid development of photogrammetry and remote sensing technology,high resolution remote sensing image has become an important geographic information data source.As an important basic geographic information,road can extract from high-resolution remote sensing data quickly and accurately,which plays an extremely important role in mapping,monitoring geographic conditions and traffic navigation.Point at the problem of low automation of existing image road extraction methods,this paper combined the genetic algorithm,the neural network,the fuzzy set theory and the mathematical morphology method and used visual theory to study road information extraction technology from high resolution remote sensing image,which realized the fast and accurate extraction and vectorization of road information and application of road network updating.The main research contents are as follows:(1)The significance of road extraction for high resolution remote sensing images,the characteristics of high resolution remote sensing images and the road characteristics of images were briefly analyzed.The research status of road extraction from high resolution remote sensing images at home and abroad was introduced,the trend of development analyzed,and a set of road extraction process based on visual theory was proposed.(2)A method for image enhancement of fuzzy sets based on improved genetic algorithm was studied.The fuzzy set enhancement algorithm can enhance the image and make the road area of the image smoother and more integrated,but the threshold is difficult to select.In this paper,a method of threshold selection based on genetic algorithm was studied,so that the selected threshold was closer to the road's grayscale feature,thus effectively enhance the characteristics of road area.The experimental results showed that the algorithm was superior to the contrast enhancement algorithm,the histogram equalization enhancement and the Ostu improved fuzzy set enhancement method in road image enhancement.(3)A new image segmentation method based on improved pulse coupled neural network(PCNN)and K-means was proposed.In view of the problem of PCNN parameter selection,the selection of internal link coefficient was proposed by the inverse distance weighting method,and the number of pixels per activation was accumulated to reach the total number of pixels as the termination iteration condition,and the KSW double threshold selection of the genetic algorithm was used as the internal activity restriction condition of PCNN.Finally,the K-means algorithm was used to further improve the segmentation effect.The experimental results showed that the improved method can improve the segmentation effect and can improve efficiency.(4)A road extraction method based on road features and multi-structure and multi-size morphology was proposed.Firstly,the geometric features of the road images were analyzed,and the structural operators of mathematical morphology were determined according to the morphological characteristics of the road.In this paper,a multi-scale and multi-structure operators were applied to image morphologic processing,which avoided the problem of weak noise immunity and loss of edge information,at the same time,ensured the integrity and accuracy of the extracted road area.Finally,the road information extracted from the road was morphologically refined,and the center line of the road was extracted.The experimental results showed that the road features based on the multi-morphological and multi-scale morphological methods were more effective.(5)Vector quantization of the extracted roads was done,and Douglas-Peucker was used to compress vectorization data,so as to reduce data storage.The MappingStar digital mapping system was used to edit and modify the road line.The extracted results were applied to the collection of road information.Compared with the traditional manual acquisition method,it can effectively improve the efficiency of collection and reduce the workload of manual editing.Finally,the road extraction method was applied to update the road network.