A new least squares method based line generalization in GIS

The generalization of line features is an important process in both traditional cartography and in many applications of geographic information systems. In this paper, the parcel boundary line generalization is further studied. The uncertainties in Douglas-Peucker based line generalization algorithm is first analyzed, a line-fitting based line generalization algorithm is then proposed. In the new line generalization method, kinds of conditional equations in map generalization are therefore derived, including the length conditional equation, area conditional equation, and end-vertexes identical equation. Based on the theoretical discussion, a parcel boundary line generalization system is developed. The case studies are carried out to test the methods. The results show that the proposed uncertainty-processing model is feasible to ensure the data quality in map generalization