This research study introduces the use of a change detection and classification algorithm that relies on the change vector analysis (CVA) method. Its implementation aims to ensure adequate response to operational production needs and allow optimized data processing over extended and environmentally complex areas. Automatic change class labelling relies on the use of a (3n+2)-dimensional feature space, where n denotes the number of sensor bands. Such enhanced feature space allows for a finer and more accurate definition of change classes of the 'from-to' type. Moreover, and to efficiently address the problem of change area overestimation, the proposed method takes into account specific evidence derived from the pixel's geographic neighbourhood, the latter defined as a 3×3 pixel kernel. The performance of this integrated algorithmic approach has been tested and validated in the framework of the CORINE Land Cover-Greece 2000 and the ESA/GSE Forest Monitoring projects in three test sites located in the outskirts of the city of Ptolemais, Thasos island and the suburbs of Athens in Greece. Its implementation in such highly fragmented and dynamically changing landscape environments has resulted in qualified and accurate land cover change maps, achieving an overall level of classification accuracy of 88-96%. Compared to visual image interpretation, the method requires half the effort. In conclusion, the proposed method has proved effective and can be recommended for use in the framework of operational projects.