In this contribution the complete implementation of an automated multiscale measurement system (AMMS) for the inspection of micro lenses and micro electromechanical systems is presented. The system uses an adaptable active exploration strategy to balance the conflict between lateral resolution, axial accuracy and measurement duration. It is equipped with several sensors with different fields of view, resolutions and accuracies. The sensors are linked flexibly during the measurement process by image processing and data fusion algorithms. Image processing algorithms are used to identify defect indicators which represent possible unresolved defects in the current sensor scale. The information, gathered by the indicator algorithms, results in new regions of interest and knowledge about the specimen feature which is needed to select and to condition more finely scaled sensors, and to trigger higher resolved measurements in the next scale. For the automated adaption and parameter optimization of the system to a measurement task, an assistant system for sensor and algorithm selection is used. We present the necessary components for automatic task adaptation and active exploration of micro lenses and micro electromechanical systems (MEMS). Inspection results for MEMS-wafer and micro lens arrays and a performance analysis are discussed.