Beam damage detection using computer vision technology

In this paper, a new approach for efficient damage detection in engineering structures is introduced. The key concept is to use the mature computer vision technology to capture the static deformation profile of a structure, and then employ profile analysis methods to detect the locations of the damages. By combining with wireless communication techniques, the proposed approach can provide an effective and economical solution for remote monitoring of structure health. Moreover, a preliminary experiment is conducted to verify the proposed concept. A commercial computer vision camera is used to capture the static deformation profiles of cracked cantilever beams under loading. The profiles are then processed to reveal the existence and location of the irregularities on the deformation profiles by applying fractal dimension, wavelet transform and roughness methods, respectively. The proposed concept is validated on both one-crack and two-crack cantilever beam-type specimens. It is also shown that all three methods can produce satisfactory results based on the profiles provided by the vision camera. In addition, the profile quality is the determining factor for the noise level in resultant detection signal.