System functions
The vision system for the greasing of ham is designed to guide an anthropomorphic robot to apply the greasing with very high precision, in a fully automatic manner.
The 3D stereo technology enables full three-dimensional reconstruction of the product surface, even with irregular shapes and complex textures.
The acquired image is divided into the main components: meat, fat and rind.
Special focus is given to the recognition of the hip, which is visually similar to the fat, and to the identification of the femur, whose three-dimensional position is derived with high accuracy.
These advanced image segmentation tasks are performed by neural networks.
Stereo RGB cameras, in single or multiple configurations, allow simultaneous acquisition of depth and color information, ensuring high image quality even on shiny or low-contrast surfaces.
The system processes the acquired information in real time, generating optimized trajectories for the greasing application, always orthogonal to the surface to be treated, even in the presence of complex curvatures.
The interface with the robot is direct, allowing autonomous and repeatable greasing, with no need for manual intervention or adjustments by the operator.
This Specialvideo vision system is on board the first and only automated system for ham sugnatura built by S.M.A. SRL, within its new division SMA-FOOD TECH.
