At the moment when the filling production line is running at high speed, the precise docking of the robot arm and the mouth of the glass bottle is like "threading a needle" - a slight shake of the bottle body or a change in ambient light may cause filling deviation. PLY-PACK's independently developed visual positioning technology is solving this industry problem with disruptive innovation. Traditional robot arms rely on preset coordinates for positioning, and the error can reach more than 3mm when facing special-shaped bottles or conveyor belt vibrations. PLY-PACK's solution is to deploy high-frame rate industrial cameras at the filling station, and build a three-dimensional point cloud model of the bottle mouth through image acquisition at 1200 frames per second. The algorithm integrates visual data with the robot arm's motion trajectory in real time, and can achieve a positioning accuracy of ±0.2mm even if the bottle is tilted 15°, which is equivalent to accurately inserting a hair on a high-speed train.
In a practical test at a jam factory in Saudi Arabia, the strong sunlight transmitted from the ceiling of the workshop caused frequent misjudgments by traditional photoelectric sensors. The PLY-PACK technical team has developed an adaptive light compensation algorithm for this purpose:
· Use multi-spectral imaging to separate ambient light interference
· Bottle mouth contour prediction model based on deep learning
· Six-dimensional force control sensor at the end of the robot arm for real-time feedback
This system enables the equipment to maintain a 99.7% positioning success rate under 1200Lux illumination fluctuations. Customer feedback said that after replacing the glass bottle mold, the production line switching time was shortened from 45 minutes to 3 minutes.
PLY-PACK uses a non-contact infrared positioning module due to concerns that contact positioning may contaminate the product. Light sources with special wavelengths can penetrate translucent liquids such as honey to capture the position of the bottle mouth without causing food ingredients to deteriorate. The German TÜV test report shows that the technology meets the compliance requirements of FDA 21 CFR Part 11, and is particularly suitable for filling highly sensitive categories such as baby food.
When the positioning error of the robot arm is compressed from millimeter level to micron level, it not only improves the filling accuracy. An application case of a dairy company shows that precise positioning improves the airtightness of the bottle cap packaging by 18%, and the shelf life of the product is extended; another seasoning company relies on the stable bottle mouth alignment effect to increase the production line speed to 240 bottles/minute. These data confirm that visual positioning technology is redefining the accuracy threshold of the packaging industry.
