The fatigue of concrete structures is related to various environmental factors and external loads such as earthquakes, wind, and other forces. The formation of cracks in concrete structures has been a significant challenge in civil engineering and has garnered considerable attention from re-searchers. Artificial intelligence has opened new horizons in estimating and predicting cracks in this context. Machine learning, facilitated by extensive databases, including crack shape, size, and the load that caused it, processes data to estimate the load that created the crack based on its shape. Specifically, crack detection begins by using standard edge detection filters with appropriate ac-curacy to extract the crack shape. This research combines manual crack detection processes with machine vision algorithms and analyzes images taken from the shear walls of two buildings at the University of Kurdistan. For this purpose, the Otsu and Canny methods are employed for crack detection. In addition, studying crack patterns and determining properties of cracks, such as length, width, and angle, using standard equations and mathematical representations of cracks, are also used to understand their nature and cause. Combining the investigated methods is ultimately ap-plied to concrete structures considered real-world examples. The results obtained from these samples provide valuable information. The analysis results for these samples show that all observed cracks result from poorly executed concrete casting processes and thermal changes.
