Automatic Recognition Of Hatched Residential Areas In Topographic Maps

At present,despite lots of digital raster maps(DRG)are stored in the mapping department,they are rarely used so that they cannot play the role effectively.Residential areas are not only one of the most important geographical elements in the map,but an important place where people live and produce.In order to make an effective assessment of the regional land conditions,some commercial insurers urgently need to identify and extract the hatched residential areas from the historical raster maps,they compare the identified results with the up-to-date map data and analyze their differences.In view of this demand,it is of great practical significance and scientific significance to carry out the research on automatic identification and extraction of residential areas.In order to automatically,efficiently and accurately recognize the hatched residential areas from raster maps.A new method based Gabor filter according the feature of hatched residential area was presented in this paper.This method can solve the current problems that likely identification errors or low accuracy of the boundary position.Firstly,the Gabor filter,Gaussian smoothing,binarization,corrosion,and logic operations were used to detect the hatched line of the hatched residential areas.Then,a boundary point on the polygon of the resident was tracked by tracking along the hatched line.Next,the boundary initial point was used as the starting point to track the next adjacent boundary point according to the relationships between the pixel values of the eight neighborhoods.In the tracking process,the relationships between the neighborhood pixel values and the designed bar detector were used to remove the noise.In this way,the boundary points were tracked in sequence,and the boundary map of the inhabitants was obtained.The boundary of the inhabitants was stored in the form of vector coordinate strings.Finally,the residential areas identification experiments of three typical areas were carried out.The experimental results showed that the method proposed in this paper may accurately and automatically identify the hatched residential areas in the raster map,and the boundaries of recognition were complete.