Context and objectives
The implementation of the European agricultural policy requires timely and accurate information about the surface areas of agricultural land use and more specifically of the economically important crops and of the crops for which subsidies are granted according to cultivated area. Each member country of the European Community collects and analyses its agricultural statistics according to a different method, which makes it difficult to compare the data on a European level. It is in this context that the Regional Inventory project MARS (= Monitoring Agriculture with Remote Sensing) was set up. Its main objective is to develop a standardised method for obtaining objective agricultural statistics with known accuracy before the end of the agricultural season.
The inventory involves a ground survey (collection of reference data) as well as remote sensing.
The use of satellite imagery allows a more efficient and accurate estimation of the acreages.
Expected scientific results
Agricultural statistics (acreage by crop) with a known level of accuracy available months earlier than the conventional NIS (National Institute of Statistics) data obtained via questionnaires
Maps of agricultural land use (1:400.000 covering Belgium, and 1:100.000 for a limited zone)
The crop acreages derived from the classification of the satellite images did not differ substantially from the results obtained from the ground survey. Substantial differences however do exist between the statistics of MARS and of the NIS. One of the main reasons is that the NIS crop acreage statistics are based on questionnaires which are filled in by the farmers.
The global classification accuracy was rather low, because : (a) the acquisition date of the satellite images was not optimal, (b) the parcel size is typically small in Belgium, and (c) only a monotemporal image set was available for most of Belgium.