Synapse quality - Sensure

FASTER, SMARTER, AND MORE EFFICIENT QUALITY CONTROL AND PACKAGING PROCESSES

In industries like food production, where product variability and external variables and fluctuations are high, ensuring consistent quality control is a constant challenge. Although quality control was initially implemented for end-product inspection, businesses are also increasingly recognizing its value in earlier stages of production. By integrating vision inspection throughout the process, manufacturers can achieve comprehensive product monitoring and improve handling and packaging efficiency like never before.

Sensure

Sensure

AUTOMATE QUALITY CONTROL
PROCESS WITH AI

Discover more

SENSURE solutions combine Artificial Intelligence and advanced machine vision technologies to automate quality control process. By leveraging three complementary approaches, the SYNAPSE software can address different inspection challenges across a wide range of products and production environments.

SELF LEARNING for Quality Control

AI-driven algorithms automatically learn product features and detect defects, irregularities, and deviations from the expected quality standards.
The system can classify products as compliant or non-compliant and automatically generate quality recipes and tolerance thresholds, reducing the need for manual setup or extensive training datasets.
Operators can easily review and adjust parameters to optimize the inspection process.
DEEP LEARNING for Ingredients and Toppings Detection

Deep learning models enable the detection and quantification of specific elements such as toppings, ingredients, or distribution patterns.
This approach is particularly effective for complex visual features that are difficult to detect using traditional vision techniques, ensuring reliable inspection even in highly variable products.
TEMPLATE MATCHING for Low-Variability Products

For products with limited variability, template matching compares items to a reference model, detecting minimal deviations in shape, structure, or surface.

Sensure

Main quality features measurable with SYNAPSE

Shape and Size (min/max/average diameter, perimeter, length, width, roundness, area, etc.), compactness, height/thickness (min/max/average), cracks/holes and their conformity, flatness, planarity, volume, slope/symmetry
Colour on top (base colour, spot detection, etc.), and/or bottom/lateral surfaces (base colour, burnt areas, flour marks, spot detection, foreign colour anomalies in general), edge/contour defects
Topping/seeding conformity (quantity, distribution, and colour), texture analysis
Foreign body detection
Temperature
Moisture

Possibility to develop ad-hoc quality features upon feasibility assessment.

Rejection mechanisms

Choosing the right rejection mechanism is important step to ensure non-compliant products are properly removed from the line even at very high speed. The choice depends on several factors, including type of products (dough, baked, with/without topping, etc.) and dimensions, line width and throughput rate/speed, products pattern required (before and after rejection), where non-compliant products need to be directed (conveyor, bin, etc.), space available, etc.

Some examples of the different rejection mechanisms include:

Conveyor shutdown or Alarm – The conveyor stops or an alarm is triggered when a non-compliant product is detected.
Air jet with nozzles – Air jets (top, bottom, or lateral) used to reject small to medium products; stronger or multiple nozzles can also remove larger items.
Traps – Conveyor-based rejection systems that temporarily lower or raise a section of the conveyor to drop non-compliant products; can be linear, hinged, or configured as drop nose bars.
Retractable belts – End rollers retract to create a gap in the product flow so items can fall through; the roller then quickly returns to position. Suitable for multi-lane applications.
Pusher – A pneumatic cylinder with a plate that pushes medium to heavy products off the conveyor; ideal for evenly spaced and oriented items.
Robot – Robotic pick-and-place systems provide high flexibility and can handle products of different sizes by changing grippers, but require more space on the production line.