7 Międzynarodowa Konferencja INŻYNIERIA PRODUKCJI – r. Wrocław Piotr Garbacz INTEGRATED VISION SYSTEMS VERSUS CUSTOM SOLUTIONS FOR QUALITY INSPECTION IN INDUSTRY
7 Międzynarodowa Konferencja INŻYNIERIA PRODUKCJI – r. Wrocław 2/14 Image analysis and processing software Automatic line integration Kick-out station, PLC, network Image acquisition triggering, synchronisation Object preparation positioning, surface cleaning System configuration lighting, optics and sensor size Industrial Vision Systems Machine Vision – multidisciplinary engineering
7 Międzynarodowa Konferencja INŻYNIERIA PRODUKCJI – r. Wrocław 3/14 Industrial Vision Systems Classification o smart camera - compact structure (integrated sensor, processor, lens and lighting) - adaptability to industrial conditions - low functionality - binary result of the inspection - easy integration into production lines
7 Międzynarodowa Konferencja INŻYNIERIA PRODUKCJI – r. Wrocław 4/14 o intelligent camera - integrated processor and vision sensor - easy integration into production lines - compact structure and adaptability to industrial conditions - short boot time - configuration with use of PC o integrated vision system - optimal lighting, lens and camera type configuration - multi-camera functionality - short boot time - configuration with or without use of PC - multiple interface options Industrial Vision Systems Classification
7 Międzynarodowa Konferencja INŻYNIERIA PRODUKCJI – r. Wrocław 5/14 o PC based vision system - based on specialized libraries * knowledge of programming language is required * high flexibility (both hardware and software) - commercial module software * knowledge of programming language is not required * limited software development Industrial Vision Systems Classification
7 Międzynarodowa Konferencja INŻYNIERIA PRODUKCJI – r. Wrocław 6/14 o DIP (Digital Image Processing) - point transformation - geometric transformation - context transformation - morphological transformation - spectrum transformation* o Image analysis algorithms - image position and orientation identification - image shape and dimensions measurement - image / object identification and comparison - completeness check - surface inspection* Industrial Vision Systems Image analysis and processing
7 Międzynarodowa Konferencja INŻYNIERIA PRODUKCJI – r. Wrocław 7/14 o inspection object - cone bearing (producer: FŁT Kraśnik) - 23 rolling elements placed in the cage - reflecting metal surface - industrial conditions (surface pollution) - low frequency of inspection - binary result of inspection - statistics reports Industrial Vision Systems Automatic Optical Inspection in cone bearings production
7 Międzynarodowa Konferencja INŻYNIERIA PRODUKCJI – r. Wrocław 8/14 o research station - vision system: Keyence CV camera: Keyence CV-H035M - lens: Keyence CA-LH25 - LED Ring Light: Latab SA(White) d=102mm - lighting Control Unit: Latab PAD Keyence CV-5000 Lighting Control Unit Ring Light Camera Industrial Vision Systems Automatic Optical Inspection in cone bearings production
7 Międzynarodowa Konferencja INŻYNIERIA PRODUKCJI – r. Wrocław 9/14 o vision system: Keyence CV image processing 16 image filters (e.g. Sobel, dilation, erosion) - image analysis 18 inspection tools(e.g. Pattern Search, OCR) - 16 different camera types - light control expansion unit - camera expansion unit (up to 4) - the parallel processing architecture CV-H035M 640x480 Industrial Vision Systems Automatic Optical Inspection in cone bearings production
7 Międzynarodowa Konferencja INŻYNIERIA PRODUKCJI – r. Wrocław 10/14 o graphical user interface Stage II Edge Position window Filter: Median, Binary, Shrink Stage I Trend Edge Position window Filter: Binary o inspection algorithm Stage III Area (1 to 46) window Filter: Binary, Shrink Total time of inspection – 84 ms Industrial Vision Systems Automatic Optical Inspection in cone bearings production
7 Międzynarodowa Konferencja INŻYNIERIA PRODUKCJI – r. Wrocław 11/14 o PC based visual system - PC: Dual Core 3,2 GHz, 4 Gb RAM - Basler Scout,GigE, PoE, 640x480, 32 FPS - software: * Microsoft Visual C Express Studio * Basler Pylon SDK * Open Source Computer Vision Library Microsoft Visual Studio Initialization, GUI Image analysis and processing Basler Pylon SDK Image acquisition Open Source Computer Vision Industrial Vision Systems Automatic Optical Inspection in cone bearings production
7 Międzynarodowa Konferencja INŻYNIERIA PRODUKCJI – r. Wrocław 12/14 o inspection algorithm Total time of image processing and analysis – 32 ms cvDilate(gray1,gray1,NULL,1); cvErode(gray1,gray1,NULL,1); cvDilate(gray2,gray2,NULL,1); cvErode(gray2,gray2,NULL,1); cvErode(gray3,gray3,NULL,1); Industrial Vision Systems Automatic Optical Inspection in cone bearings production
7 Międzynarodowa Konferencja INŻYNIERIA PRODUKCJI – r. Wrocław 13/14 o systems comparison Integrated vision systems PC based vision systems Very High flexibility and functionality Difficult installation and configuration High flexibility and functionality Simple installation and configuration Knowledge of programming language is not required Knowledge of programming language is usually required Time of processing depends on algorithm complexity Time of processing depends on system-level design optimization Lower implementation costs Lower upgrade costs Industrial Vision Systems Summary
7 Międzynarodowa Konferencja INŻYNIERIA PRODUKCJI – r. Wrocław 14/14 Thank you for your attention!