| Optical camera | 5 million high-speed intelligent digital industrial camera |
| Resolution (FOV) | Standard configuration 15 μ M/Pixel (corresponding FOV: 38mm * 30mm) 10/15/20 μ M/Pixel (optional) |
| Optical lens | 5M pixel telecentric lens, depth of field: 8mm-10mm |
| Light source system | Highlight RGB coaxial ring multi angle LED light source |
| operating system | Windows 10 Professional |
| Computer configuration | I3 CPU, 8G GPU graphics card, 16G memory, 120G solid state hard disk, 1TB mechanical hard disk |
| Machine power supply | AC 220 V ± 10%, frequency 50/60Hz, rated power 1.2KW |
| Machine dimensions | 1100mm × 900mm × 1350mm （L × W × H) Height including footing |
| Machine weight | 450kg |
| Optional configuration | Offline programming software, external barcode gun MES traceability system interface open |
| Size | forty × 40mm~450 × 330mm (larger size can be customized according to customer requirements)) |
| Thickness | 0.3mm～6mm |
| Plate weight | ≤3KG |
| Clear height | Upper clear height ≤ 35mm, lower clear height ≤ 70mm (customized for special requirements) |
| Minimum test element | 0201 element, IC with 0.3mm spacing and above (01005 element can be selected) |
| Solder paste printing | Presence, deflection, less tin, more tin, open circuit, pollution, tin connection, etc |
| Part defects | Missing parts, offset, skew, stele, side standing, turnover, polarity reversal, wrong parts, damage, multiple pieces, etc |
| Solder joint defect | Less tin, more tin, connected tin, faulty soldering, multiple pieces, etc |
| Wave soldering inspection | Insertion pin, Wuxi, less tin, more tin, faulty soldering, tin bead, tin hole, open circuit, multiple pieces, etc |
| Red glue plate detection | Missing parts, offset, skew, monument, side standing, turnover, polarity reversal, wrong parts, damage, glue overflow, multiple pieces, etc |
1,Colinear algorithm: The backboard LED light strip needs to detect the relative offset between the LED and the LED, ensure that the entire LED light strip is colinear, and perfectly slove the industry problem of S-tyle non-collinear LED distribution test, and truly achieve non-adjacent LED collinear analysis.
2,Resistance Value Recognition: This algorithm used the latest machine learning technology to calculate the exact resistance by recognizing the characters printed on the reisitor. The algorithm can be used todetect the wrong parts of the resistor and automatically matchthe function of “alternative material”.
3,Scratch detection: the algorithm will search for dark stripes of a specified length in the target area and calculate the average brightness value of the dark stripe area. The algorithm can be used to detect scratches, cracks, etc. on flat surfaces.
4,Intelligent judgment: This algorithm collects various qualifiedand umfavorable image samples, establishes an intelligent judgment model throughtraining, and calculates the similarity ofthe images to be measured, This algorithm simulates the human thinking mode and is difficult to detect problems with some tradityinal algorithms.You can handle iteasily. For esample: peak welding spot detection, resetting tin ball detection, polarity detection of circular components, etc..
