The visualization and simulation industry has a demonstrated interest in classification products for sensor simulation. The challenge lies in providing highly accurate material classification of remotely sensed imagery while significantly reducing the time and cost to create products. Visualization and simulation products for material classification are created by merging and mosaicking multi-source satellite and aerial imagery of different resolutions on an elevation surface to provide realistic, geo-specific terrain features. This requires that all image data is orthorectified, seamlessly co-registered, tonally balanced and feather blended into mosaics from source data of different resolution. To achieve highest accuracy at faster speed and lower cost, we apply an innovative, optimal pixel-labeling process to the mosaic imagery. This process is based on artificial intelligence (AI) algorithms using Nash Equilibrium and game theoretic analyses to help solve the problem of feature extraction through supervised classification. This can be viewed as a constant sum game, whereby the players are pixel data points that take part in the game to decide their class memberships. A player's land cover classification strategies are based on four different supervised learning algorithms: k-Nearest Neighbors (KNN), Decision trees using a classification and regression tree (CART), Normal/Naïve Bayes probabilistic graphical model, and support vector machine (SVM). Within this formulation, we used a weighted reward matrix for consistent labeling of feature pixels and classification factors, resulting in higher accuracy and precision when compared to the individual machine learning algorithms alone.