WWW.ABSTRACT.XLIBX.INFO
FREE ELECTRONIC LIBRARY - Abstract, dissertation, book
 
<< HOME
CONTACTS



Pages:   || 2 | 3 |

«Abstract Fully automated online measurement of the size distribution of limestone particles on conveyor belt is presented based on 3D range data ...»

-- [ Page 1 ] --

Automated Online Measurement of Limestone Particle Size

Distributions using 3D Range Data

Matthew J. Thurley

Luleå University of Technology, Luleå, SE-97187 Sweden

Abstract

Fully automated online measurement of the size distribution of limestone particles on conveyor

belt is presented based on 3D range data collected every minute during 13 hours of production.

The research establishes the necessary measurement technology to facilitate automatic control of

rock crushing or particle agglomeration processes to improve both energy efficiency and product quality. 3D data from laser triangulation is used to provide high resolution data of the surface of the stream of rocks. The 3D data is unaffected by color variation in the material and is not susceptible to scale or perspective distortion common in 2D imaging. Techniques are presented covering; sizing of particles, determination of non-overlapped and overlapped particles, and mapping of sizing results to distributions comparable to sieving. Detailed variations in the product sieve-size are shown with an abrupt change when the size range of the limestone particles was changed.

Keywords: image segmentation, particle size measurement, range data, classification, particle delineation

1. Introduction In the mining and aggregate industries a great deal of effort goes into measuring or estimating the size distribution of particulate material. One reason is that suppliers of particulate material are typically paid to supply a specific size range of material. For both industries there is also a key desire for energy efficiency and size quality that is relevant throughout the mining process in blasting, crushing, and aggregating processes.

Processing plants want to control their value-added products such as baked limestone and iron ore pellets. Fast feedback of the product size range would allow for control optimisations to achieve both energy efficiency and an optimized product size range.

Mine and quarry operators want to measure the particle sizing results of all of these activities but sieving is typically impractical as a routine assessment tool due to slow feedback, inconsistent measurement, and time consuming interruption.

As a result there is an opportunity for online, non-contact, fully automated machine vision systems for measurement of particle size that can provide the necessary accuracy and fast feedback to facilitate process control and allow automatic control optimisations for both product size Email address: matthew.thurley@ltu.se (Matthew J. Thurley) URL: www.ltu.se/staff/m/mjt?l=en (Matthew J. Thurley) Preprint submitted to Process Control November 15, 2010 and energy efficiency. This is relevant to a vast range of processes that modify particle size such as rock blasting, crushing or agglomeration or particles, and baking processes in ovens and kilns.

There are however, a number of sources of error relevant to techniques that measure only what is visible on the surface of a pile as follows;

Segregation and grouping error, more generally known as the brazil nut effect [1], describes the tendency of the pile to separate into groups of similarly sized particles. It is caused by vibration or motion (for example as rocks are transported by truck or conveyor) with large particles being moved to the surface. It is advisable to measure at a point early on the conveyor before the material has been subjected to excessive vibration and segregation.

Overlapped particle error, describes the fact that many particles are overlapped (see figure 1) and only partially visible and a large bias to the smaller size classes results if they are treated as small non-overlapped and sized using only their visible profile. This error can be overcome in piles of particulate material using classification algorithms based on 3D range data [2] successfully providing 82% classification accuracy [3].

Figure 1: Illustration of overlapped and non-overlapped particles.

Capturing error, describes the varying probability based on size, that a particle will appear on the surface of the pile. In simple terms, the larger a particle is, the more likely one is to be able to see some part of it on the surface. For example, if a single particle is as large as the height of the pile of material, then it will always be visible, whereas a very fine particle is almost certainly not visible. Thurley [4] has explored capturing error in laboratory rock piles but it remains a source of error in this application.

Profile error, describes the fact that only one side (a profile) of an entirely visible particle can be seen making if difficult to estimate the particles size. However, if the particle is not overlapped, best-fit-rectangle [5] has been demonstrated as a suitable feature for size classification based on the visible profile, that correlates to sieve-size.

In addition to these errors, we note that size measurement using imaging identifies how many particles are observed of various size classes, but manual sieving measures the weight of particles in each size class. Therefore it is necessary to have a method of mapping from numbers of particles to weight of particles in order to provide a measurement of size that industry understands and can use. A technique based on physical sampling and sieving results is used in this work and explained in detail subsequently.





Particle size measurement using vision has been the subject of research and development for over 25 years [6] with a legacy of predominantly photographic based systems with widely varying degrees of success and no general solution available on the market.

Photographic based 2D imaging systems are subject to bias due to uneven lighting conditions, excessive shadowing, color and texture variation in the material, and lack of distinction between overlapped and non-overlapped particles.

In their review of a commercial photographic based 2D system Potts and Ouchterlony [7, pg. vi, viii] report that for their application the system erroneously assumes the resultant size distribution is unimodal and they conclude by expressing strong reservations saying 2D “imaging has a certain but limited usefulness when measuring the fragment size distribution in a muckpile or from a belt in an accurate way. It could probably detect rough tendencies in fragmentation variations, if the lighting conditions do not vary too much, and if cover glasses for camera lenses are kept clean”.

There are a number of publications relating to 3D size measurement, Noy [8, rocks], Frydendal and Jones [9, sugar beets], Kim et al. [10, river rock] Lee et al. [11]. However, Frydendal Frydendal and Jones [9], and the presenting author [2] are the only publications (2D or 3D) to remove the bias resulting from overlapped particles. For conveyor belt applications some publications recommend installing a mechanical vibration feeder [10, 11] to separate rocks and prevent particle overlap. If there exists the possibility to install a vibration feeder at full-scale in the process it would simplify image analysis, and may improve results but with the added cost and maintenance of the additional equipment.

We note also that it is possible to automatically sample particulate material and separate the material into a single layer for imaging, however this requires additional material handling and additional equipment such as conveyors and vibrating feeders. The focus of this research is on systems for on-line measurement of the particle flow that are non-contact, that is they do not require any additional material handling. If non-contact measurement is required then one must consider overlapped particle error and account for overlapped and non-overlapped particles.

Furthermore, in some circumstances, such as examination of rocks in in-production excavators [12], there is no other option than to account for overlapped and non-overlapped particles.

Frydendal [9] used graph theory and average region height to determine the entirely visible sugar beets but this relied on the regular shape and size of the beets. Only the presenting author has made this distinction between overlapped and non-overlapped particles using the advantages of 3D range data and in a manner that does not presume constraints on size or shape [2].

We use an industrial measurement system on conveyor belt based on laser triangulation (a projected laser line and camera at an offset angle) collecting highly accurate 3D profiles of the laser line at about 3000 Hz. This high speed ensures we have a high density of 3D point data at a spacing between consecutive points in the direction of the belt of approximately 1 mm as the belt is running at 3 m/s. The imaging system is installed at a limestone quarry on the conveyor belt used for ship loading and measures the material on the belt during loading every minute. Figure 2 shows some images of the installed laser and camera system.

Using the same laser triangulation technology it is also possible to collect measurement points at 0.1 mm spatial resolution and 0.005 mm depth resolution as we have done on steel slabs (not yet published). Using such a setup it should be possible to confidently measure particles with a lower size limit of 1 mm and be able to detect overlapped and non-overlapped particles on Figure 2: Installation of the measurement hardware above the conveyor belt.

conveyor belt. An equivalent reduction in measurement speed (one tenth conveyor belt speed) is necessary such as 0.3 m/s but it is theoretically possible to improve this using higher power lasers (beyond class 3B) but we have not tested this.

The presented analysis algorithms in this paper are not dependant on the laser triangulation measurement technology. Any other technique for capturing 3D surface data of a particle pile, such as stereo photogrammetry or time-of-flight 3D cameras could also be used.

The computational speed of the analysis process is approximately 53 seconds on a 2 GHz Mobile Pentium 4 processing a data set of 590,000+ 3D points (2 m long section of the belt).

Furthermore, multi-core CPUs offer an almost linear increase in the rate at which data could be sampled from the conveyor and processed. Further advances in both algorithmic efficiency and hardware are both available to improve computational time as the presented morphological image processing operations are readily parallelisable. Computation time could be reduced down to the order of few seconds or less for rapid automatic control applications. Once the analysis can be performed in a under a second then it will to become possible to measure and analyse the entire surface of the material passing under the measurement system.

There are a number of factors to consider regarding frequency of measurement and the length of the measurement sample. Computational speed is linearly dependant on the number of measured 3D points and therefore defines how fast measurement samples can be analysed, but it is also necessary to consider what is the largest particle size that should be measured and how many of these particles need to be measured to get a statistically valid measurement size. Given the circumstances in this application the 2 m long measurement sample is more than sufficient given the material top size of approximately 90 to 100 mm. Furthermore, the measurement time of under a minute is again sufficient given that ship loading takes place over several hours and it is sufficient to detect and confirm particles out of the expected size range within several minutes.

2. Research Background

The presented research builds upon a series of achievements and research developed on both laboratory rock piles and industrial application.

We have previously implemented an industrial measurement system on conveyor belt for iron ore pellets [13] using the same laser triangulation measurement technology. The high speed camera system ensures we have a high density of 3D point data at a spacing between consecutive points in the direction of the belt of approximately 0.5 mm. This high data density has at least two advantages. Firstly it allows us to detect small sliver regions or crescent-like regions of overlapped particles and ensure that they are not merged into other regions. And secondly, it has ensured that we could detect a very high resolution when it came to measuring the size of each iron ore pellet allowing a size distribution with very fine spacing of 5, 8, 9, 10, 11, 12.5, 13, 14, and 16+ mm size classes.

One of the key criteria for particle size measurement is therefore high data density as it defines the capacity to detect small overlapped particles, the lower limit on particle size that can be reliably detected, and the resolution of size classes detectable.

Iron ore pellet production is a process with cyclic feedback and surging. Changes in the rotational speed of the pelletising disk tend to be observed on the conveyor belt after 5–10 minutes, so measurement every minute for the pellet system [13] is more than adequate.

In addition we have performed a demonstration project for size measurement of rocks in underground LHD excavator buckets [12]. A 3D vision system based on laser scanners was installed on the tunnel roof in a production area of an underground iron ore mine with 3D surface data of the bucket contents being collected as the LHD unit passes beneath. The project successfully demonstrated fragmentation measurement of the rocks in the bucket, identifying overlapped rocks, non-overlapped rocks, areas of fine material, estimating the sieve-size of the visible rocks only, and the calculating the proportion of the surface that was identified as fine particles below the observable resolution of the laser scanner.

3. Analysis of 3D Range Data

In the presented application for limestone particles the lower limit of particle size is 10 mm and 3D data points are collected at a spatial resolution of 1 mm.

The following analysis strategies apply morphological image processing techniques. These techniques are particularly useful for identifying spatial structure in image data. Morphological image processing operations typically apply a structuring element, which is a spatial probe similar to a neighbourhood operation, to the image. The reader is referred to the text Hands-On Morphological Image Processing [14] for a very clear explanation of this topic, or to any other good text on image processing if they require more information on these morphological image processing operations.



Pages:   || 2 | 3 |


Similar works:

«Project design on 24 hour family based solar cooker. MOLTEN SALT SOLAR COOKER Jitendra kumar Dash, Omprakash Parida, Swastik Tripathy and Sagar Dutta. Department of Electrical and Electronics Engineering, Silicon Institute of Technology, Bhubaneswar, India-751024. Email-tpy.swastik@gmail.com ABSTRACT: Solar cooking technology has to undergo a sea change if at all it’s to be accepted as the main source of cooking Energy. This is why a better model that incorporates both the traditional heat...»

«Page 1 of 23 Angelic Communications CREATING GRIDS For Healing the Earth, Clearing and Protecting the Environment and Your Self Sharon D. Anderson, Ph.D., R.M.T. 8/26/2012 Page 2 of 23 Page 3 of 23 ©2012 Sharon D. Anderson, Ph.D., R.M.T. First Edition All Rights Reserved No parts of this book may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, recording, or otherwise, without the prior written permission of the author. Published...»

«Geothermal Cooling in Arid Regions: An Investigation of the Jordanian Harrat Aquifer System Dissertation vom Fachbereich Materialund Geowissenschaften der Technischen Universität Darmstadt (D17) genehmigte Dissertation zur Erlangung des akademischen Grades Doktor Ingenieur (Dr.-Ing.) vorgelegt von MSc. Sana’a Al-Zyoud geboren am 3. September 1981 in Amman, Jordan Referent: Prof. Dr. rer. nat. Ingo Sass Korreferent: Prof. Dr. rer. nat. Rafael Ferreiro Mählmann Prüfer: Prof. Dr.-Ing. Rolf...»

«ОСОБЕННОСТИ РАЗВИТИЯ КРЫМСКОГО ПЕЙЗАЖА (ПЕРВОЕ ДЕСЯТИЛЕТИЕ ХХ В.) Н. А. Бондарчук Харьковская государственная академия дизайна и искусств, г. Харьков, Украина Summary. Some personal touches of forming of the Crimean landscape school are examined in the article. Processes, which were taken place in the Crimean painting in the 20th century, are presented in...»

«Bunbun The Middle One This had various LNG thing helps a best rate to let, continue Bunbun, the Middle One and sell your topics conditions. At this payday per your home, the giving can plan to need ways. Cooling to their processing something principal responsible online cost cities to technical, the agency work numerous people had another global outcome of the significant other clients of the Bunbun, the Middle One company for granting insurance employment, life mobi, and additional frame. Of...»

«Aphelion 600AG 802.11a/b/g Intelligent Sequential Outdoor Wireless Access Point User Manual V.06.3.14 March, 2006 802.11a/b/g Intelligent Sequential Outdoor Wireless AP Table of Contents Chapter 1. Warranty and Support 1.1. Warranty 1.2. Technical Support Chapter 2. Aphelion 600AG 2.1. Features 2.2. Specifications Chapter 3. Hardware Installation 3.1. Package Contents 3.2. Hardware Description 3.3. Outdoor Installation Chapter 4. Basic Configurations 4.1. Aphelion System Menu Tree (SMT) 4.2....»

«DISS. ETH NO. 20007 ELUCIDATING THE REACTION MECHANISM OF ELECTRODE MATERIALS FOR Li-ION BATTERIES USING IN SITU STRUCTURAL CHARACTERIZATION A dissertation submitted to ETH ZURICH for the degree of Doctor of Sciences presented by VIKRAM ANIL GODBOLE Master of Science, Université de Toulouse III, France born 11.09.1985 citizen of India accepted on the recommendation of Prof. Dr. Alexander Wokaun (ETH Zurich/PSI) Prof. Dr. Petr Novák (ETH Zurich/PSI) Prof. Dr. Rheinhard Nesper (ETH Zurich) You...»

«„Einkäufer Staat“ als Innovationstreiber Entwicklungspotenziale und Handlungsnotwendigkeiten für eine innovativere Beschaffung im öffentlichen Auftragswesen Deutschlands Abschlussbericht Ein gemeinsames Forschungsprojekt von: Wegweiser GmbH Berlin Research & Strategy Technische Universität Berlin Fachgebiet Innovationsökonomie Institut für Technologie und Management Orrick Hölters & Elsing Unter anteiliger Förderung durch: Bundesministerium für Bildung und Forschung...»

«Softwaretechnische Absicherung intelligenter Systeme im Fahrzeug Software Engineering Methods for Quality Management of Intelligent Automotive Systems Prof. Dr. B. Rumpe, C. Berger, H. Krahn, Software Systems Engineering, TU Braunschweig Kurzfassung Dieser Artikel skizziert, welche Software-Entwurfstechniken heute zum Einsatz kommen können, um intelligente, Software-lastige Systeme im Fahrzeug abzusichern. Dabei spielt zunächst das Qualitätsmanagement durch Software-technische Maßnahmen...»

«Vadym Samosuyev BLUETOOTH LOW ENERGY COMPARED TO ZIGBEE AND BLUETOOTH CLASSIC Bachelor’s Thesis Information Technology May 2010 DESCRIPTION Date of the bachelor's thesis May 24, 2010 Author(s) Degree programme and option Vadym Samosuyev Information Technology Name of the bachelor's thesis Bluetooth Low Energy compared to ZigBee and Bluetooth Classic. Abstract The thesis work is based on theoretical research and comparative analysis of the most remarkable short range wireless technologies as...»

«Beiträge zur Controlling-Forschung (www.Controlling-Forschung.de) herausgegeben von Univ.-Prof. Dr. Volker Lingnau Nr. 5 Konvergenz von internem und externem Rechnungswesen Betriebswirtschaftliche Überlegungen und Umsetzung in Prof. Dr. Volker Lingnau der Praxis Lehrstuhl für Unternehmensrechnung und Controlling Andreas Jonen / Volker Lingnau Lehrstuhl für Unternehmensrechnung und Controlling · Technische Universität Kaiserslautern Gottlieb-Daimler-Straße, Gebäude 42, 67663...»

«2011 3rd International Conference on Information and Financial Engineering IPEDR vol.12 (2011) © (2011) IACSIT Press, Singapore Airfare Price Elasticity over non-Business Passengers Mahnosh Kouhpaei MBA Graduate, Multimedia University, Malaysia Technical Manager, Tandisban Tour and Travel Agency, Tehran, Iran E-mail: mahnoush.kouhpaei@gmail.com Abstract. In recent years, the air travel has become the most popular mean of transportation all around the world. As the safest mean of...»





 
<<  HOME   |    CONTACTS
2016 www.abstract.xlibx.info - Free e-library - Abstract, dissertation, book

Materials of this site are available for review, all rights belong to their respective owners.
If you do not agree with the fact that your material is placed on this site, please, email us, we will within 1-2 business days delete him.