One primary detractor in an alignment system’s performance is the corruption of alignment target integrity by thin film processes, metal grain structure, and contamination. While most aligners will acquire even very low contrast signals with sufficient precision, the measured align position is often inaccurate due to process induced asymmetries or dislocation of edges in the targets. If marks are sufficiently corrupted by lithographic processes, all alignment systems will suffer a performance degradation albeit to different degrees depending on the design. This article will present a general methodology and algorithm for the real-time detection and correction of erroneous alignment signal data in an image forming alignment system. An alignment signal figure of merit, based on image symmetry and internal consistency [A. Stankov, A. Lupata, and C. Progler, Proceedings of the KTI Microelectronics Seminar, City, Data 1991 (unpublished)], forms the basis for judging the integrity of alignment data. These two alignment signal quality metrics are computed and checked during each alignment cycle. Based on the signal metrics, the algorithm permits special signal filtering to obtain more robust alignment performance. Moreover, the algorithm has the ability to quantify the "quality" of alignment targets prior to exposure, thus, highlighting potential problem layers, mask errors, and wafer trouble spots due to processing. A description of the algorithm logic and implementation is described along with test bed and exposure-based data demonstrating the many benefits of the algorithm. These include improved overlay, faster alignment convergence, fewer alignment fails, and rapid assessment of alignment target integrity saving both time and money in the early stages of process development. In addition, an extension of the concept to a broader range of alignment system designs is briefly described.