Machine Vision System

We need a Machine Vision System (MVS) that uses four to twelve cameras and can template-match the Complete Knit Pattern (CKP) of a garden hose as it exits a Knitting Machine (KM). Our stitch pattern uses a radial wrap laid beneath a plain stitch knit (Diagram, “4 Spiral Wraps” & Image, “Knitted 0.625 Tube”). The MVS shall halt the KM when a variation in the CKP is detected. A simple output to signal our PLC to halt the KM shall be sufficient. Variation in the CKP is typically caused by a breakage of one of the twelve needles resulting in an absence of one of the twelve needle rows. When a needle breakage occurs the only variation in the CKP will be the absence of one needle row. The MVS shall ensure the presence of the following within the CKP: 1. Four radial wraps (Diagram, “4 Spiral Wraps”) 2. Twelve needle rows (Diagram, “8 Courses of Knit”) 3. Eight courses per inch (Diagram, “8 Courses of Knit”) Other details and remarks: • Product to be inspected: clear tubing with white polyester reinforcement knit over the outer surface • Tubing size: 3/4” OD, 5/8” ID • Tubing rate-of-speed: 36 +/- 0.01 FPM • The knitting process is very consistent—the knit does not rotate or jump • Preferred hardware for MVS: o Raspberry Pi o Pi Camera o Arduino Deliverables shall include all the following: 1. MVS that includes a machine vision program able to capture and compare the live subject with a fixed standard image (template matching). The template shall be able to be captured and loaded via the inspection camera(s) via a “subject capture” mode. 2. BOM for the MVS 3. Wiring diagram for the MVS 4. Complete access to programming and programming documentation of the MVS 5. Included technical consultation until the MVS is working