The truck will transport the pallet full of materials to the unloading area, and after unloading by forklift truck, it will be directly transported to the 3D manipulator station for sorting and unstacking. After completion, AGV will be used to cooperate with the transportation to the corresponding work station.
Share
1). Project background:
Customer demand:
Identify and disassemble different size bins.
Customer pain points:
1. different sizes of the box mixed identification, according to the position and size of the box output corresponding grabbing direction.
2. The exposure changes in real time from morning to night (sunny or cloudy), and the recognition rate of the product and the stability of the system need to be improved.
2).Solution architecture:
Overall process:
Key process modules:
Character comparison: Set the corresponding matching command, which is used to logically trigger subsequent module detection.
Workpiece grab: Draw the area of the RGB image where the workpiece is to be grabbed, with the scope of the box selection beyond the outer edge of the actual frame.
Condition detection: Determine whether the product is recognized, and send corresponding OK/NG logic signals to the host computer.
Multi-size box: configure the relevant parameters of the box palletizing type of a variety of sizes and carry out the task of unpalletizing, use the example to segment the algorithm model, collect pictures of a variety of product placement positions for model training.
2). Program advantages:
1. Under the premise of large exposure changes, the image can be clear and the recognition is stable, and the time from the received signal to the output data is 3.5s.
2. for a variety of sizes, colors of the material box to achieve 100% recognition rate.
3. to achieve 24/7 uninterrupted production, capacity efficiency increased by 100%.
4. a 3D-RGBD mechanical manual station can replace 5 manpower.
EN
