Image processing system to detect temporal changes in outside environments

This papers deals with the problem of automatic temporal change detection in outside environments. Because of the limitations of conventional approaches, we developed an image processing system based on attributes deduced from the Walsh-Hadamard spatio-frequential transformation. Taking into account perturbations due to the outside environment, image enhancements and a background modelisation were added to guarantee the robustness of the system. Results are provided in the context of an industrial application.