Image-based crack assessment of concrete structures using artificial intelligence for bridge health monitoring

Concrete crack, which signifies the decrease of the structural reliability and integrity, is commonplace. Human-based visual inspection is currently involved to address the development of the surface cracks. However, this practice is highly subjective and unreliable. Through the research work, an image-based algorithm for concrete crack detection and quantification methods was developed. Moreover, experiments were conducted to verify the performance of the proposed FCN algorithm and to convert the measurement scale from pixels to millimetres, which in turns will facilitate a meaningful structural evaluation process.