Application of Adaptive Anisotropic Diffusion Filter To Improve the Reflector Image Quality when Performing Ultrasonic Nondestructive Testing

At present, systems for the recognition (classification) of reflectors based on artificial neural networks are being actively developed to increase the speed of preparation of the ultrasonic testing protocol and reduce the influence of the human factor. For these systems to operate most efficiently, the reflector images must be processed in order to increase the signal-to-noise ratio of the image and its segmentation (clustering). One way of segmentation is to process the image with an adaptive anisotropic diffusion filter, which is used to process optical images. The effectiveness of using this texture filter for segmentation of reflector images reconstructed from echo signals measured with antenna arrays has been demonstrated in model experiments.