Ultrasonic internal defect detection in cheese

Different ultrasonic signals and detection techniques were used and compared to detect internal foreign bodies present in semi-soft cheeses. The signals were a pulse or a chirp and the detection was carried out by using either correlation with a reference signal or a wavelet decomposition. The principle of the detection consisted in measuring the time of flight of the transmitted signals and of the echoes, the latter in the absence of foreign body should be the double of the former. The presence of a foreign object affected this pattern in several ways. In order to assess the method, a small plastic cylindrical object of 3 mm in diameter was introduced in one half of the cheese and was tested for detection, the second half being used as reference for the control cheese. The results showed that the two signals and the two detection methods were able to localise the transmitted signals and the echo from the opposite face of the cheese under all circumstances. For the foreign body detection, the correlation method gave superior results, in term of signal to noise ratio as well as in term of error rate, while the two signals gave similar results. The analysis of the mean and standard deviation of the signal to noise ratio of the object echo showed that some samples presented peak values close to those due to the noise. Nevertheless, the object was detected in 90% of the tests. There was no significant effect of temperature on the detection technique.