PROJECT TITLE :
A Novel Graph-Matching-Based Approach for Domain Adaptation in Classification of Remote Sensing Image Pair
This paper addresses the problem of land-cover classification of remotely sensed image pairs in the context of domain adaptation. The primary assumption of the proposed technique is that the training knowledge are obtainable only for one in every of the photographs (source domain), whereas for the other image (target domain), no labeled knowledge are offered. No assumption is made here on the amount and therefore the statistical properties of the land-cover classes that, in flip, may vary from one domain to the opposite. The only constraint is that a minimum of one land-cowl category is shared by the 2 domains. Below these assumptions, a unique graph theoretic cross-domain cluster mapping algorithm is proposed to detect efficiently the set of land-cover classes which are common to each domains as well as the additional or missing classes within the target domain image. An interdomain graph is introduced, which contains all of the category info of each images, and subsequently, an economical subgraph-matching algorithm is proposed to spotlight the changes between them. The proposed cluster mapping algorithm initially clusters the target domain knowledge into an optimal range of teams given the on the market source domain coaching samples. To the present finish, a method based mostly on information theory and a kernel-primarily based clustering algorithm is proposed. Considering the actual fact that the spectral signature of land-cowl categories may overlap significantly, a postprocessing step is applied to refine the classification map produced by the clustering algorithm. Two multispectral knowledge sets with medium and very high geometrical resolution and one hyperspectral data set are thought-about to judge the robustness of the proposed technique. Two of the data sets comprises multitemporal image pairs, whereas the remaining one contains pictures of spatially disjoint geographical areas. The experiments ensure the effectiveness of the proposed framework in several complex eventualities.
Did you like this research project?
To get this research project Guidelines, Training and Code... Click Here