The Novel Artificial Intelligence Based Sub-Surface Inclusion Detection Device and Algorithm

Abstract

We design, implement, and test a novel tactile elasticity imaging sensor to detect the elastic modulus of a contacted object. Emulating a human finger, a multi-layer polydimethylsiloxane waveguide has been fabricated as the sensing probe. The light is illuminated under the critical angle to totally reflect within the flexible and transparent waveguide. When a waveguide is compressed by an object, the contact area of the waveguide deforms and causes the light to scatter. The scattered light is captured by a high resolution camera. Multiple images are taken from slightly different loading values. The distributed forces have been estimated using the integrated pixel values of diffused lights. The displacements of the contacted object deformation have been estimated by matching the series of tactile images. For this purpose, a novel pattern matching algorithm is developed. The salient feature of this sensor is that it is capable of measuring the absolute elastic modulus value of soft materials without additional measurement units. The measurements were validated by comparing the measured elasticity of the commercial rubber samples with the known elasticity. The evaluation results showed that this type of sensor can measure elasticity within ±5.38 %.