Agriculture classification using POLSAR data

Growing crops display a wide range of canopy geometries and shapes of plant components. From the radar point of view, this means that different crops distribute the dielectric material of which they are made differently in space: their architectures vary a lot. Some crops (or at least some of their components) show strongly preferred orientations, such as the stalks or ears of cereals. The importance of SAR polarimetry in crop classification arises principally because polarisation is sen-sitive to orientation. Hence it provides a means to distinguish crops with different canopy archi-tectures. Detailed explanation of how this occurs is not straightforward, since the polarimetric response is determined by both attenuation and scattering processes, and these in turn depend on the probing frequency and the incidence angle as well as plant properties. In addition, the scattering mechanisms can vary with depth in the crop canopy, particularly between the response of the canopy itself and soil response. It is expected that PolInSAR data will add to the classification potential of POLSAR data by their sensitivity to the vertical distribution of scatterers. Different approaches have been used to classify SAR data, and a very important class of algorithms is the knowledge-based approaches. Here, generic characteristics of different cover types are derived by combining physical reasoning with the available empirical evidence. These are then used to define classification rules. Because of their emphasis on the physical content of the SAR data they attempt to generate robust, widely applicable methods, which are nonetheless capable of taking local conditions into account. In this paper a classification approach is presented, that uses a knowledge-based approach, where the crops are first classified into broad classes, i.e. bare surfaces, cereal crops, root crops, spring crops, winter crops etc. Hereafter, the discrimination between the individual crop types within these broad categories is performed. The discrimination into the broad classes is performed using standard polarimetric parameters such as the backscatter coefficients both at linear and circular po-larisations, correlation coefficient and phase difference. The classification into the individual crop types is possible for some of the crop types, whereas it may for instance be very difficult to discriminate between some of the cereal crops. This part of the classification process is not as well established as the first part, and both a supervised approach and a knowledge-based approach have been evaluated. Both POLSAR and PolInSAR data may be included in the classification scheme. The classification approach has been evaluated using data from the Danish EMISAR and the German ESAR, and quantitative results based on in-situ information will be presented.