Abstract
A decision-making helping tool was set up by the integration of GIS and AHP for peri-urban agriculture planning in Hanoi province (Vietnam). This process is based on a raster representation of the study area. So the accuracy depends of the pixel size, but also, with today high resolution imaging, it can be difficult to maintain a synthetic understanding of the global spatial distribution of different kinds of urban and rural areas. It happens especially when they are interwoven for creating a specific landscape, only characterized by patterns or colours distribution. In order to have an helpful tool for decision makers, we applied Papri texture analysis techniques for improving the final map of peri-urban agriculture planning in Hanoi province while preserving the best possible accuracy for key locations (roads, river banks, downtown...). Papri is a method for texture analysis that can be applied both on quantitative datas (colour images) or on qualitative datas (classifications, maps). Also Papri is working with multiple size neighborhoods, so it can preserve a good accuracy by using small neighborhoods (border zones) and it can also make a good generalization in the heart of each landscape. We show in this study that the overall accuracy after simplification by texture analysis is maintained with the same range of values for peri-urban areas high, medium, low and not suitable for peri-urban agriculture, while the ability in understanding the new spatial phenomenons is greatly improved.