Abstract: Currently, most cases of land adaptability evaluation are based on either raster data model or vector model, boiling down to the incompleteness of information in the evaluation process and costing the assessment accuracy. Therefore, the combination and the full use of the advantages of raster and vector data models together could largely improve the accuracy and efficacy in land adaptability assessment. Based on such idea as objectoriented modeling, this article puts forward vectorraster mixed data model for land adaptability evaluation as well as the approaches of implementation. Firstly, raster data is segmented to raster objects by the linear boundary of land unit in vector data, and then the raster objects are linked to their corresponding vector counterparts with unique raster object labels. Next, statistic features in each raster object are extracted and added to the vector counterparts as attributes. Following this approach, information of each land adaptability index is figured out and stored in the database for future calculations. At last, according to the weights of each index, the comprehensive evaluation is conducted by weighted summation for each land unit. To demonstrate the improvement of this method, the constructions of high quality prime farmland in Zhejiang Province is taken as a case study.

Key words: vector, raster, mixed data model, land adaptability evaluation, HQPF