Nonlinear adaptive threshold for ultrasound image process

Ultrasound reflection tomography (URT) is a novel technique which has been evaluated for flow imaging and measurement in multiphase component flowing mixtures through the real-time reconstruction of a cross-sectional image. This paper presents a nonlinear threshold which aims to enhance the reconstructed ultrasound tomographic image dynamically. Because of the finite beam profile for the ultrasonic transducer employed, the resultant signal intensity in a reconstructed 2-dimensional grid is characterised as a nonlinear distribution function, therefore, a conventional linear threshold is not able to process such a heterogeneous image for optimal contrast. Furthermore, the signal to noise rate (SNR) in a reconstructed image varies depending on the scattering, size and number of insonified objects, therefore, so a statically scaled threshold is not suitable for this situation either. This paper demonstrates the application of a nonlinear adaptive threshold, which is able to produce tomographic images with improved contrast and better consecutive object boundaries.