Agricultural Land cover mapping by active learning from multispectral spot-7 satellite image

International audience Agricultural practices are major drivers of water flows in cultivated landscapes. Land Covermapping have a strong impact onto run off and soil erosion at the landscape and watershedscales. In practice, there is a need for agricultural land cover mapping at both scales thatcould be integrated into hydrological models to better understand the considered hydrologicalbehavior and thus improve water resources managementField methods are inadequate for characterizing the spatial variability of crops at such scales.Remote sensing appears therefore as a promising alternative. The classification of agriculturalland cover using a single multispectral satellite images has proven to be challenging due to thehigh parcel intra variability over a wide study area.Our objective is land cover mapping of agricultural fields, at a large scale, using active learningtechniques. The methods are applied to a spot-7 multispectral satellite image over a 35 km2 areaof Lebna catchment in the North Eastern Tunisia. The image high spatial resolution allows alarge scale land cover mapping on a wide extent.The proposed method is based on a supervised classification, using field observations aslearning samples. The first difficulty consists is having a very small number of learningsamples. The classifier model is then constructed locally, making it suboptimal for the entirearea. Indeed, the learning samples are assumed to be representative of the whole data set whichis rarely confirmed in practice on a wide area.To solve this issue, the proposed method is based on active learning techniques that build anefficient learning set by improving iteratively the model performance by adding the mostinformative samples. Samples are then labelled by the user and allow to construct a newclassifier model that should be more efficient. The selection of new samples needs a strategy torank pixels. Two criteria are used and often coupled: uncertainty and diversity. The samplesshould be the most informative (I.e uncertain for the current classifier model) and diverse (I.enon redundant).The used uncertainty measure is based on a random forest classifier probability. This measurecan be applied to any other classifier that provides a probability of belonging to differentclasses. For sample diversity, two metrics are used; a similarity-based and clustering-basedtechniques.Besides, this paper focuses on an operational strategy that allows mapping agriculture cover ona large wide area. Two schemes are possible; either a pixel-based classification or a parcel-based classification. The latter assumes having a digitized parcels; using GIS techniques, of thestudy area but has the advantage of highly reducing computing times and leading to a land covermap at the parcel scale that can be directly used in hydrological models.First results show the effectiveness of active learning techniques to map a wide area whilemaintaining a small number of training samples.