Ultrasound offers a versatile imaging modality. It has a long history of medical use, but system and signal processing limitations have limited its full range of capabilities. By examining other uses of ultrasound, such as materials testing, the effectiveness of medical ultrasound tissue identification could likely be improved. To a limited degree, low frequency ultrasound has been used examine the texture of carrots tissue. These studies focused on the velocity and attenuation of low frequency ultrasound through a carrot sample; these results yielded a degree of ambiguity when attempting to identify a carrot texture. In our research, we attempted to characterize the texture of a carrot, based upon its frequency response. Research showed that there are a number of sources of variation when imaging a carrot. First, the response of the transducer itself can vary. Replications of the same sample also yield varied results. This is a result of problems such as transducer misalignment, or coupling. Imaging different segments of a carrot can be a significant source of error. By far, the largest source of variation came from the imaging samples from different carrots. This variation is partly due to different transducer coupling, and possible non-uniform cooking. Also, to an unknown extent, slight differences in cell biology among different carrots also contributes to this high variation. A response look up table was compiled for carrots cooked .5 to 16 minutes. When an attempt was made to identify the cooking times of 10 unknown samples, 4 of them were correctly identified. This inaccuracy in matching is due to high sources of variation noted above. A more customized matching program could yield better results. By combining frequency response studies with velocity and attenuation studies, a more accurate characterization model could likely be devised.
Vick, Christopher, "High frequency ultrasonic characterization of carrot cell texture" (1999). Thesis. Rochester Institute of Technology. Accessed from
RIT – Main Campus