Intelligent Robots for Handling of Flexible Objects. IRFO Vision System

Similar documents
Robot Vision: Camera calibration

A 100Hz Real-time Sensing System of Textured Range Images

Reconstruction of Deformation from Depth and Color Video with Explicit Noise Models

Task analysis based on observing hands and objects by vision

Rigid ICP registration with Kinect

Computed Photography - Final Project Endoscope Exploration on Knee Surface

Shape Modeling with Point-Sampled Geometry

Structured Light II. Thanks to Ronen Gvili, Szymon Rusinkiewicz and Maks Ovsjanikov

Computer Graphics 1. Chapter 7 (June 17th, 2010, 2-4pm): Shading and rendering. LMU München Medieninformatik Andreas Butz Computergraphik 1 SS2010

Object Classification in Domestic Environments

fmri pre-processing Juergen Dukart

: Easy 3D Calibration of laser triangulation systems. Fredrik Nilsson Product Manager, SICK, BU Vision

3D Computer Vision. Structured Light II. Prof. Didier Stricker. Kaiserlautern University.

Improving the 3D Scan Precision of Laser Triangulation

Ceilbot vision and mapping system

Processing 3D Surface Data

Global Non-Rigid Alignment. Benedict J. Brown Katholieke Universiteit Leuven

3D Digitization of Human Foot Based on Computer Stereo Vision Combined with KINECT Sensor Hai-Qing YANG a,*, Li HE b, Geng-Xin GUO c and Yong-Jun XU d

CALIBRATION BETWEEN DEPTH AND COLOR SENSORS FOR COMMODITY DEPTH CAMERAS. Cha Zhang and Zhengyou Zhang

A consumer level 3D object scanning device using Kinect for web-based C2C business

ICP and 3D-Reconstruction

Processing 3D Surface Data

A Study of Medical Image Analysis System

High-Fidelity Augmented Reality Interactions Hrvoje Benko Researcher, MSR Redmond

Dynamic Rendering of Remote Indoor Environments Using Real-Time Point Cloud Data

Surface Registration. Gianpaolo Palma

Lecture Week 4. Images

Miniature faking. In close-up photo, the depth of field is limited.

Structured light 3D reconstruction

3D graphics, raster and colors CS312 Fall 2010

3D Object Representations. COS 526, Fall 2016 Princeton University

Mobile Point Fusion. Real-time 3d surface reconstruction out of depth images on a mobile platform

Camera Calibration for Video See-Through Head-Mounted Display. Abstract. 1.0 Introduction. Mike Bajura July 7, 1993

Inside VR on Mobile. Sam Martin Graphics Architect GDC 2016

3D Models from Range Sensors. Gianpaolo Palma

Capture of Arm-Muscle deformations using a Depth Camera

CSL 859: Advanced Computer Graphics. Dept of Computer Sc. & Engg. IIT Delhi

SC: Prototypes for Interactive Architecture

Polygonal Mesh. Geometric object made of vertices, edges and faces. Faces are polygons. Polyhedron. Triangular mesh Quad mesh. Pyramid Cube Sphere (?

Depth Buffer Based Registration of Free-form Surfaces

Registration of Dynamic Range Images

TEXTURE OVERLAY ONTO NON-RIGID SURFACE USING COMMODITY DEPTH CAMERA

Efficient Grasping from RGBD Images: Learning Using a New Rectangle Representation. Yun Jiang, Stephen Moseson, Ashutosh Saxena Cornell University

This one-semester elective course is intended as a practical, hands-on guide to help you understand digital art.

And. Modal Analysis. Using. VIC-3D-HS, High Speed 3D Digital Image Correlation System. Indian Institute of Technology New Delhi

Gesture Recognition: Hand Pose Estimation. Adrian Spurr Ubiquitous Computing Seminar FS

MULTI-MODAL MAPPING. Robotics Day, 31 Mar Frank Mascarich, Shehryar Khattak, Tung Dang

3D Perception. CS 4495 Computer Vision K. Hawkins. CS 4495 Computer Vision. 3D Perception. Kelsey Hawkins Robotics

WITH A KEEN EYE FOR QUALITY AND COST

Learning Semantic Environment Perception for Cognitive Robots

Tecnologie per la ricostruzione di modelli 3D da immagini. Marco Callieri ISTI-CNR, Pisa, Italy

Nonlinear State Estimation for Robotics and Computer Vision Applications: An Overview

Project Title: Welding Machine Monitoring System Phase II. Name of PI: Prof. Kenneth K.M. LAM (EIE) Progress / Achievement: (with photos, if any)

Specification and Computation of Warping and Morphing Transformations. Bruno Costa da Silva Microsoft Corp.

Registration of Moving Surfaces by Means of One-Shot Laser Projection

Nonrigid Surface Modelling. and Fast Recovery. Department of Computer Science and Engineering. Committee: Prof. Leo J. Jia and Prof. K. H.

Fast Tracking of Deformable Objects in Depth and Colour Video

Acquisition and Visualization of Colored 3D Objects

DEPTH ESTIMATION USING STEREO FISH-EYE LENSES

Prof. Fanny Ficuciello Robotics for Bioengineering Visual Servoing

Fast 3D Reconstruction and Recognition

Overview. Augmented reality and applications Marker-based augmented reality. Camera model. Binary markers Textured planar markers

CALIBRATION ASPECTS OF MULTIPLE LINE-SCAN VISION SYSTEM APPLICATION FOR PLANAR OBJECTS INSPECTION

HIGH SPEED 3-D MEASUREMENT SYSTEM USING INCOHERENT LIGHT SOURCE FOR HUMAN PERFORMANCE ANALYSIS

Sculpting 3D Models. Glossary

A Study on the Distortion Correction Methodology of Vision Sensor

Discuss Proven technologies that addresses

Graphics Systems and Models

Introduction to Mobile Robotics

L16. Scan Matching and Image Formation

MARK. January 2019 issue... BENCH THE INTERNATIONAL MAGAZINE FOR ENGINEERING DESIGNERS & ANALYSTS FROM NAFEMS

PERFORMANCE CAPTURE FROM SPARSE MULTI-VIEW VIDEO

For 3CCD/3CMOS/4CCD Line Scan Cameras. Designed to be suitable for PRISM based 3CCD/CMOS/4CCD line scan cameras

This week. CENG 732 Computer Animation. Warping an Object. Warping an Object. 2D Grid Deformation. Warping an Object.

Augmented Reality, Advanced SLAM, Applications

Rich Augmented Reality Interactions (without goggles, gloves or 3D trackers)

3D Vision Real Objects, Real Cameras. Chapter 11 (parts of), 12 (parts of) Computerized Image Analysis MN2 Anders Brun,

Rectification and Distortion Correction

Measurements using three-dimensional product imaging

3D Modeling of Objects Using Laser Scanning

Three-dimensional nondestructive evaluation of cylindrical objects (pipe) using an infrared camera coupled to a 3D scanner

Dedicated Software Algorithms for 3D Clouds of Points

MR-Guided Mixed Reality for Breast Conserving Surgical Planning

Optimized Design of 3D Laser Triangulation Systems

Image Processing using LabVIEW. By, Sandip Nair sandipnair.hpage.com

Data Visualization. What is the goal? A generalized environment for manipulation and visualization of multidimensional data

Assessing Accuracy Factors in Deformable 2D/3D Medical Image Registration Using a Statistical Pelvis Model

Full Field Displacement and Strain Measurement. On a Charpy Specimen. Using Digital Image Correlation.

3D Computer Vision 1

Synchronization aspects of sensor and data fusion in a research multi-sensor-system

Real-time Image-based Reconstruction of Pipes Using Omnidirectional Cameras

Human-body modeling Xi et al. [3] used non-numerical attributes like gender, age, race, marital status, and occupation to predict body shapes and buil

CS 395T Numerical Optimization for Graphics and AI (3D Vision) Qixing Huang August 29 th 2018

International Conference on Communication, Media, Technology and Design. ICCMTD May 2012 Istanbul - Turkey

TEXTURE OVERLAY ONTO NON-RIGID SURFACE USING COMMODITY DEPTH CAMERA

An introduction to 3D image reconstruction and understanding concepts and ideas

Efficient SLAM Scheme Based ICP Matching Algorithm Using Image and Laser Scan Information

Removing Moving Objects from Point Cloud Scenes

1 Introduction Motivation and Aims Functional Imaging Computational Neuroanatomy... 12

Image Transformations & Camera Calibration. Mašinska vizija, 2018.

Transcription:

Intelligent Robots for Handling of Flexible Objects IRFO Vision System Andreas Jordt Multimedia Information Processing Institute of Computer Science University Kiel

IRFO Vision System Overview 2) Sensing of Material Properties Estimating Material Parametes by Haptic / Visual Sensing Tasks: Object Scanning Manipulation Density Simulation Visual Sensing by Range Camera 1) Sensing of object shape Physical Modelling Shape Prediction 3) Modeling and Predicion Execution Learning 4) Manipulation and Learning Object Tracking 2

IRFO Vision System Overview Depth/Color Sensors Scanning Stage Interaction Stage Conveyor 3

Scanning Stage Depth/Color Sensor: Kinect 640 x 480 Pixel @ 30Hz Framerate Simultaneous Color and Depth Video 4

Scanning Stage: 3D Model Generation Task: Acquire a 3D model Generate 3D surface from depth image Remove conveyor surface (Optional): Colorize 3D surface 5

Scanning Stage Deformation Estimation Grasp calculation Database Deformation Tracking 6

Deformation Tracking Task: Determine Object Deformation Define deformation model Find model parameters describing the deformation 7

Deformation Tracking: Deformation Model NURBS Surfaces A continuous surface function defined by a set of 3D control points. 8

Deformation Tracking: Deformation Model Define a vector positions. NURBS Surfaces, which contains all 3D control point Register the 3D mesh from the scanning stage to such a NURBS surface. 9

Registration Consider every 3D point of the given model Find the closest point on the surface for each vertex Describe every object point relative to the NURBS x d z y u v 10

1 1

With a deformable 3D model at hand, how do we determine the correct deformation parameters for every frame in a give sequence? Render deformation guesses and compare them to the input data Define a fitness function to measure the difference between the results of a deformation and the input images.. 12

The Fitness Function How is such a function defined? Color fit (normalized luminance) Depth fit Surface penalty (Regularization) 13

Fitness function evaluation Controlpoints C For every object point: Recursive NURBS evaluation Transformation to D Projection into D Apply lense distortion Compare to depth value C Transformation to C Projection into C W D Apply color lens distortion Normalize and compare color Evaluate distance points Calculate distances 14

Fitness function evaluation D C Build NURBS Cache Transform CPs Undistort Images Warp C into D Pre-normalize Color For every object point: Recursive NURBS evaluation Transformation to D Projection into D Apply lense distortion Compare to depth value Transformation to C Projection into C Apply color lens distortion Normalize and compare color Evaluate distance points Calculate distances 15

CMA-ES Optimization problem has e.g. 48 dimensions for 4 x 4 control points High dimensional optimization problem with no derivation available => Covariance Matrix Adaptation (Evolution Strategy) Samples (Individuals) per Iteration: 4 + floor ( 3 * log(dim) ) e.g. 10 for dim=100 16

18

19

20

Thanks Thank you for your attention. 21