THE VISIONLAB TEAM engineers - 1 physicist. Feasibility study and prototyping Hardware benchmarking Open and closed source libraries

Similar documents
23270: AUGMENTED REALITY FOR NAVIGATION AND INFORMATIONAL ADAS. Sergii Bykov Technical Lead Machine Learning 12 Oct 2017

Lecture 19: Depth Cameras. Kayvon Fatahalian CMU : Graphics and Imaging Architectures (Fall 2011)

ME 6406 MACHINE VISION. Georgia Institute of Technology

Hochperformante Inline-3D-Messung

Measuring intensity in watts rather than lumens

Active Stereo Vision. COMP 4102A Winter 2014 Gerhard Roth Version 1

INNOVATIVE SPECTRAL IMAGING

Applying Automated Optical Inspection Ben Dawson, DALSA Coreco Inc., ipd Group (987)

SMART LASER SENSORS SIMPLIFY TIRE AND RUBBER INSPECTION

Job Description. Commitment: Must be available to work full-time hours, M-F for weeks beginning Summer of 2018.

NELA Brüder Neumeister GmbH

A NEW NEUROMORPHIC STRATEGY FOR THE FUTURE OF VISION FOR MACHINES June Xavier Lagorce Head of Computer Vision & Systems

Model-Based Design for Sensor Systems

TEXTILE INSPECTION INDUSTRY OVERVIEW

Dr Antony Robotham - Executive Director

Machine Vision for the Life Sciences

Introduction to Machine Vision for New Users

FSI Machine Vision Training Programs

AI Application Processing Requirements

Intelligent interaction

CAPACITIES FOR TECHNOLOGY TRANSFER

Considerations: Evaluating Three Identification Technologies

velociraptor HS Velociraptor is fast running and fast grabbing! Save a tree...please don't print this document unless you really need to.

ARMY RDT&E BUDGET ITEM JUSTIFICATION (R2 Exhibit)

nanovea.com PROFILOMETERS 3D Non Contact Metrology

High-speed Micro-crack Detection of Solar Wafers with Variable Thickness

Practical Image and Video Processing Using MATLAB

Advances in slabs defects inspection with Conoscopic Holography. Ignacio Alvarez, J.M. Enguita (UniOvi) J. Marina, R.

Fein. High Sensitivity Microscope Camera with Advanced Software 3DCxM20-20 Megapixels

3D-scanning system for railway current collector contact strips

Superfast phase-shifting method for 3-D shape measurement

Face Detection System on Ada boost Algorithm Using Haar Classifiers

Advanced Mechatronic System For In-Line Automated Optical Inspection Of Metal Parts

Digital Photogrammetry. Presented by: Dr. Hamid Ebadi

Confocal NEXIV VMZ-K Series. CNC Video Measuring System CONFOCAL NEXIV. VMZ-K Series

Vision with Precision Webinar Series Augmented & Virtual Reality Aaron Behman, Xilinx Mark Beccue, Tractica. Copyright 2016 Xilinx

Neural Networks The New Moore s Law

Introduction. Lighting

Fast, flexible, highly reliable image acquisition

Lessons learned & Future of FeedMAP

MacroPATH The new line of Digital Imaging Systems for Grossing

Omni-Directional Catadioptric Acquisition System

CSI: Rombalds Moor Photogrammetry Photography

e2v Launches New Onyx 1.3M for Premium Performance in Low Light Conditions

High Resolution Spectral Video Capture & Computational Photography Xun Cao ( 曹汛 )

Exercise questions for Machine vision

Computer Vision. Howie Choset Introduction to Robotics

More Info at Open Access Database by S. Dutta and T. Schmidt

3DUNDERWORLD-SLS v.3.0

Confocal NEXIV VMZ-K Series. CNC Video Measuring System CONFOCAL NEXIV. VMZ-K Series

Implementation of Augmented Reality System for Smartphone Advertisements

New Paradigm in Testing Heads & Media for HDD. Dr. Lutz Henckels September 2010

Introduction to Computer Vision

DATAVS2 series.

Embedding Artificial Intelligence into Our Lives

preface Motivation Figure 1. Reality-virtuality continuum (Milgram & Kishino, 1994) Mixed.Reality Augmented. Virtuality Real...

ADAS COMPUTER VISION AND AUGMENTED REALITY SOLUTION

THICKNESS CHECK NON-CONTACT THICKNESS MEASUREMENT OF FLAT PRODUCTS

Hyper-spectral, UHD imaging NANO-SAT formations or HAPS to detect, identify, geolocate and track; CBRN gases, fuel vapors and other substances

truepixa Chromantis Operating Guide

CNC Video Measuring System NEXIV VMZ-K series. CNC Video Measuring System. Confocal Model

Controlling vehicle functions with natural body language

Applied Machine Vision

CMOS X-RAY DETECTORS. for Industrial Non-Destructive Testing COMMITTED TO PEOPLE. DRIVEN BY INNOVATION.

DATAVS2 series.

Vision & Industry 4.0: Towards smarter sensors. Dr. Amina Chebira Vision Embedded Systems, CSEM SA October 4 th, 2016

Product Introduction

Industry 4.0. Advanced and integrated SAFETY tools for tecnhical plants

A machine vision system for scanner-based laser welding of polymers

WHO. 6 staff people. Tel: / Fax: Website: vision.unipv.it

An Electronic Eye to Improve Efficiency of Cut Tile Measuring Function

switzerland Commission II, ISPRS Kyoto, July 1988

Image Processing Architectures (and their future requirements)

Perception. Read: AIMA Chapter 24 & Chapter HW#8 due today. Vision

INNOVATION+ New Product Showcase

IN-SIGHT 2000 VISION SENSORS

flexible lighting technology

ScrappiX. Visual inspection equipment for dimensional. and surface defects control

Evaluating Barcode Reading Technologies

A NOVEL VISION SYSTEM-ON-CHIP FOR EMBEDDED IMAGE ACQUISITION AND PROCESSING

XM: The AOI camera technology of the future

Augmented and Virtual Reality

COMPUTER. 1. PURPOSE OF THE COURSE Refer to each sub-course.

Image Processing Architectures (and their future requirements)

Ricoh's Machine Vision: A Window on the Future

TL2 Technology Developer User Guide

Industrial Keynotes. 06/09/2018 Juan-Les-Pins

Acuity. Acuity. AP620 Profile Measurement Scanner. Principles of Operation. Definitions. Typical Scanner Applications

DAVID SLS-1. Getting Started Guide

OPTIV CLASSIC 321 GL TECHNICAL DATA

Technology trends in the digitalization era. ANSYS Innovation Conference Bologna, Italy June 13, 2018 Michele Frascaroli Technical Director, CRIT Srl

Basler. Line Scan Cameras

PLazeR. a planar laser rangefinder. Robert Ying (ry2242) Derek Xingzhou He (xh2187) Peiqian Li (pl2521) Minh Trang Nguyen (mnn2108)

SINGLE AND MULTIPLE CAMERA SOLUTIONS FOR AUTOMATED MACHINE VISION APPLICATIONS

Hyperspectral imaging (HSI) goes embedded All rights reserved Max Larin, 1

INSPECTION SENSORS SVS2 SERIES VISION SENSORS HIGHLIGHTS APPLICATIONS

Multi-application platform for education & training purposes in photonical measurement engineering & quality assurance with image processing

Camera Overview. Digital Microscope Cameras for Material Science: Clear Images, Precise Analysis. Digital Cameras for Microscopy

Depth estimation using light fields and photometric stereo with a multi-line-scan framework

DEEP LEARNING ON RF DATA. Adam Thompson Senior Solutions Architect March 29, 2018

Transcription:

VISIONLAB OPENING

THE VISIONLAB TEAM 2018 6 engineers - 1 physicist Feasibility study and prototyping Hardware benchmarking Open and closed source libraries Deep learning frameworks GPU frameworks FPGA programming Augmented Reality frameworks CAD Tools 2

COLLABORATIONS University of Bologna 1 PhD student Graduate candidates Computer vision projects image S Feasibility study and prototyping Hardware benchmarking 3

PROJECTS Computer vision: Blister inspection Textile defect inspection Volume measurement Caps inspection Spectral classification OCR Image Stitching Camera calibration 4

PROJECTS Deep Learning pedestrian detection people tracking object recognition 2D/3D face recognition handwritten OCR 5

PROJECTS Augmented Reality AR4ACD: assembly/repair assistance remote assistance marker based application cad based application 6

THE VISION LAB IDEA Test benches for feasibility studies Enhance the collaboration with the University of Bologna Strong partnership with Italian leader supplier (image S) in computer vision HW-SW Training of technical personnel Real perception of Computer Vision advantages Visibility to the industrial world 7

THE VISION LAB DEEP LEARNING PHOTOMETRIC STEREO GENERAL PURPOSE STEREO AUGMENTED REALITY LINE SCAN CAMERA LASER TRIANGULATION 8

STEREO VISION - ENSENSO The stereo camera IDS Ensenso N20 ( Ensenso) combines two matricial cameras and one projector in order to create 3D images of the scene. The Ensenso camera uses the same approach of a standard stereo camera with the addition of a projector that overlaps a high contrast texture to the scene so as to obtain a more homogeneous reconstruction of all the surfaces. 9

STEREO VISION - ENSENSO Resolution 1280 x 1020 Spectrum Color Focal length 6-16 mm f-number 1.6 Dimensions 175 x 50 x 50 mm Mass 550 gr Power consumption 9.5 W 10

STEREO VISION - ENSENSO Handling Material feeding, assembly Volume measurement Luggage, Parcels, Dimensional weight Quality Control Checking assemblies, blisters Security Access control Tracking people movement Medical Shep of body parts, custom splints Head movement during tomography 11

PHOTOMETRIC STEREO The photometry technique is based on the acquisition of several images of the same object subject to different illumination source. Using this technique it is possible to get information abount the analyzed surface. The matricial camera Teledyne DALSA Genie Nano M1280 Mono ( Teledyne DALSA Inc.), paired with a ring light CCS HPR2-150SW ( CCS Inc.), makes possible an optimal and accurate analysis. 12

PHOTOMETRIC STEREO Resolution 1280x1024 Spectrum Grayscale Focal length 8 mm f-number 1.4 Dimensions 21 x 29 x 44 mm Mass 46 gr Power consumption 5 W (PoE) 13

PHOTOMETRIC STEREO Surface inspection Illumination correction Embossed OCR Blister inspection 3D surface reconstruction 14

LASER TRIANGULATION - GOCATOR The Gocator ( LMI TECHNOLOGIES INC.) is a 3D sensor that can collect depth information about the analyzed scene. The 2300 series allows to create a high precision tridimensional model, that can be used for many applications, like Pick and Place, quality control and surface inspection. A conveyor is used so as to get a realistic approach to the problem. 15

LASER TRIANGULATION - GOCATOR Resolution Z 0.019-0.060 mm Resolution X 0.275-0.550 mm Clearance Distance 300 mm Measurement range 400 mm Field of View 158-365 mm Dimensions 49 x 75 x 272 mm Mass 1300 gr Power consumption 13 W 16

LASER TRIANGULATION - GOCATOR Electronics Automotive Dimensions measurement Pick and place Rubber and Tire Gap inspection, welded edge measurement Automation Coplanarity checks of chip pins, height inspection of critical electronic components Tire geometry measurement Wood Rapid multi-point 3D profiling Strips measurement 17

LINEAR CAMERA The Teledyne Dalsa Linea Color 2k line scan camera ( Teledyne DALSA Inc.) makes it possible to acquire 3-channels lines of pixels with a resolution of 2048 pixels per line. The Linea Color camera, thanks to its small size and a high speed acquisition rate (up to 45 lines per second thanks to the TurboDrive technology), offers excellent performances, especially when paired with a conveyor belt. 18

LINEAR CAMERA Resolution 2048 x 2 Spectrum Color Focal length 25 mm Bit depth 12 bit Max line rate 45 khz Dimensions 62 x 62 x 47 mm Mass 280 gr Power consumption 7W 19

LINEAR CAMERA Recycling Food inspection Material inspection Pharmaceutical labels inspection Textile inspection 20

GENERAL PURPOSE The general purpose system is composed by several modular components that can be arranged in different configurations. The workstation is currently provided with a BMT BIBL-E-IR220/110H backlight ( BMT Industrial Ltd.), two CCS SL-200X15R light bars ( CCS Asia Pte Ltd.), a Dome BMT BIDL-E-W150 illuminator ( BMT Industrial Ltd.) and a RGB Teledyne DALSA Genie Nano C1920 camera ( Teledyne DALSA Inc.). Dalsa camera Dome Bar light Backlight 21

GENERAL PURPOSE Resolution 1936x1216 Spectrum Color Focal length 16 mm f-number 1.4 Dimensions 21 x 29 x 44 mm Mass 46 gr Power consumption 5 W (PoE) 22

GENERAL PURPOSE Illumination and lens test Object measurement Quality control OCR 23

GENERAL PURPOSE The general purpose system is composed by several modular components that can be arranged in different configurations. The workstation is currently provided with a BMT BIBL-E-IR220/110H backlight ( BMT Industrial Ltd.), two CCS SL-200X15R light bars ( CCS Asia Pte Ltd.), a Dome BMT BIDL-E-W150 illuminator ( BMT Industrial Ltd.) and a RGB Teledyne DALSA Genie Nano C1920 camera ( Teledyne DALSA Inc.). Dalsa camera Dome Bar light Backlight 24

DEEP LEARNING Pedestrian detection People tracking Object recognition Handwritten OCR Depth/Semantic segmentation 25

DEEP LEARNING Algorithm acceleration CNN training CNN inference Image processing Video decoding Applications Industrial Embedded vision Video surveillance... 26

AUGMENTED REALITY The ODG R-7 smartglasses ( 2018 Osterhout Design Group) with a 2.4 GHz quad-core processor and 3 GB of RAM allow to smoothly execute heavy AR applications. The applications are easy to use thanks to two 720p displays and onboard control keys. These smartglasses support voice control so as to allow the user to work hands free and to be completely focused on the operations. 27

AUGMENTED REALITY CPU Qualcomm Snapdragon 2.7GHz quad-core RAM 3 GB LP-DDR3 Storage 64 GB Battery 1300 mah Cameras 1080p @ 60 fps 720p @ 120 fps Field of View 30 Mass 180 gr 28

AUGMENTED REALITY Industrial applications Medical and health Games and entertainment E-commerce, retail GPS drive/walk navigation 29

THE VISIONLAB Test benches for feasibility studies Real perception of Computer Vision systems Training of technical personnel Collaboration with the University of Bologna Strong partnership with image S in computer vision HW-SW 30

T3LAB Via Sario Bassanelli, 9/11 Bologna (BO) ITALY Phone +39-051-58.70.187 Contact point Luca Piccinini luca.piccini@t3lab.it - +39-051-58.70.183 31