The nozzle cleaning, detection and sorting machine adopts 3 × 5-megapixel cameras for detection. The camera field of view is approximately 29×22, with a single-pixel precision of about 0.01mm, enabling high-precision measurement. It is composed of a set of special visual measurement software.
The machine detects in two stations (divided into upper and lower stations). Operators place the small fixtures loaded with products in fixed positions and press the start button. The upper-station camera first moves to check whether each product is placed in position one by one, then confirms if the large fixture has empty slots ready. After completion, the gripper moves to transplant the products from the small fixture to the large fixture. After air drying and pressure testing, the lower-station camera takes photos to detect the appearance and dimensions of the products, finally giving an OK/NG result for each product.
Quality: The nozzle condition cannot be measured numerically.
The cleaning steps and judgment criteria are vague.
Time: Huge workload in nozzle reassembly and cleaning operations.
Cost: No man-hours allocated for nozzle maintenance.
Detect product appearance and dimensions. A total of four product types.
Y-axis
Operators place the small fixtures loaded with products in fixed positions and press the start button.
The upper-station camera first moves to check whether each product is placed in position one by one, then confirms if the large fixture has empty slots ready.
Gripper
Large fixture
After completion, the gripper moves to transplant the products from the small fixture to the large fixture. After air drying and pressure testing, the lower-station camera takes photos to detect the appearance and dimensions of the products, finally giving an OK/NG result for each product.
Air drying
Pressure testing
Lower-station camera
Small fixture ①
Small fixture 2
Small fixture 3
Camera
Product
Lens
Light source
Light source
Lens
Product
Camera
Upper station
Lower station
X-axis
Z-axis
Upper-station camera
This project uses 3 × 5-megapixel cameras for detection, with a camera field of view of approximately 29×22 and a single-pixel precision of about 0.01mm, enabling high-precision measurement. It is equipped with a set of special visual measurement software.
This detection scheme uses two stations (divided into upper and lower stations):
Front Side | Back Side
Character position, whitening, blackening detection. | 3. Outer ring radius
Distance | 4. Radius
| 5. Radius
| 6. Presence of central hole
| 7. Radius
Front Side | Back Side
Radius | 1. Radius
Radius | 2. Radius
Length and width | 3. Radius
Radius | 4. Radius
Radius | 5. Radius
Radius | 7. Radius
Front Side | Front Side
Radius | 3. Character position, whitening, blackening detection
Radius | 4. Cylindrical radius
Radius | 6. Radius
| 8. ...
Back Side | Front Side
7. Radius | 1. Radius
8. Width | 2. Radius
9. Radius | 5. Radius
10. Appearance: Deep scratches and cuts | 6. Radius
| 8. ...
| Back Side | Front Side
| Severe deformation | Nozzle damage
| Detectable | Detectable
| Nozzle missing | Slight deformation
| Detectable | Undetectable