An approach for objective and automated identification of pork belly firmness


Uttaro, B., Zawadski, S., Juárez, M. (2020). An approach for objective and automated identification of pork belly firmness, 169

Plain language summary

Pork bellies are variable, and some are more suitable for the production of side bacon than others. Suitability depends strongly on how soft or firm a belly is. To date, this determination has been carried out subjectively, by hand, at a research level, by draping a ribbed belly over a rod or bar and measuring how much it has bent after 1-2 minutes.
A series of three studies was performed to determine if an objective belly firmness assessment method could be developed for automated on-line use in production plants, that was faster than the research method. Taken into account were: being able to sort bellies very early in the carcass fabrication process; and the impact of flattening on the rib-in belly, a technique used on some production lines.
The result was that the firmness could be easily and quickly assessed when bellies were fed, ham end first, over the edge of a moving conveyor held at either a 0° or 30° incline. Manually flattening the belly had slightly more impact on the amount of bend realized on the 0° than on the 30° conveyor.
Further work is required to determine the impact of conveyor speed, how different amounts of firmness could be categorized, and the relationship of this method to the current research method.


A series of studies was performed to develop and test a method adaptable for early automated sorting of pork bellies based on firmness. Flattened and non-flattened, bone-in, skin-on primal bellies were fed skin-down, caudal end foremost on to a horizontal (0°) or raised (30°) conveyor with an adjustable nosebar (ø = 14 mm). The drop angle, after 24 cm of belly had passed the nosebar, was strongly correlated with subjective floppiness (r = 0.77–0.82; P ≤.0001) and moderately correlated with fat thicknesses (r = 0.47–0.67; P ≤.0001). On a 0° conveyor, drop angle relationships were generally weakest for non-flattened bellies, but moderate and similar for flattened bellies at 0°, as well as for both flattened and non-flattened bellies at 30°. The method appears to show promise for commercial production use. Further work is required on the impact of belt speed, firmness categorization, and the relationship to the current bar bend research method.

Publication date