Tel.+49 30 838 62550 Fax.+49 30 838 46029
The state of the art in ultrasonic detection of abscesses was confirmed: Without an advanced image analysis technology there is only a 90% probability of abscess detection, restricted to abscess diameters of minimum 0.5 cm and with measuring and evaluation times of about 1 minute.The research performed has demonstrated the feasibility of an online-method to detect abscesses in pork necks with ultrasound.Like for all automated measurements, evaluations and interpretations - to reach this goal, a strict freeze of the boundary conditions specific for the application is necessary.Under the following circumstances a nearly 100% probability of detection can be reached:1 . Real-time ultrasound system with B-scan mode 2. Working frequency of 5 MHz, brightness resolution 256 grey levels 3. Working temperature > 6"C 4. Definition of specific working path for the automatic evaluation 5. Length of the working path min 25 cm 6. Image acquisition speed about 3 images per second and 7. About 1 cm between two probepositions for image acquisition 8. Measuring place at the weigher of the slaughter-house 9. Automated image pre-analysis ref. to disturbances like air-layers (i-L spherical shape) fibre structures (# spherical structure) 10. Automated image pre-analysis by median-filter for noisesuppression 11. Automated image analysis with sliding brightness threshold and evaluation threshold at about 25% of the maximum brightness 12. Automated image analysis with "horizonal erosion" ref. to spherical up to elliptical shape 13. Abscess detection - limit at 0.5 cm 0 minimum, corresponding to 500-1800 pixelsUnder these conditions an evaluation and interpretation time of 10 seconds per pork neck can be obtained during online operations.The research has shown too, that smaller abscess dimensions cannot be detected free of doubt by ultrasound due to their geometry and the structure of their surrounding tissue. The use of higher frequencies would not improve this because of the interdependence between a higher resolution and a lower penetration depth.The application potential for this system consists for example in the completion of the offical meat inspection and in the operational quality control.