Industrial Applications

Introduction

The remainder of this page gives an outline and introduction to some of the industrial applications that have been tackled by various members of the group.

Contents

1. Rock Size Measurement
2. Bottletop Seal Inspection
3. Fruit Quality Inspection
4. Flotation Cell Surface Froth Analysis

Rock Size Measurement

Researcher

Robert Crida

Sponsor

MINTEK, Randburg

Dates

1992 to 1995

Synopsis

Gold ore must be milled in order to grind it down to a fine powder to facilitate the extraction of the gold. This is a costly process and directly affects the yield of gold from the ore. It has been proposed that a measure of the size distribution of rock material in the mill can be used to improve the control of the milling process which could then result in lower costs and improved yields.

This research was aimed at performing an on-line measurement of the size distribution of rock material on the mill's feed conveyor belt and using that measurement to estimate the distribution of the contents of the mill. Image processing has been proposed as a solution for this application since it offers several advantages in the harsh environment of a mineral processing plant:

• A non-intrusive measurement so that there is no contact with the rock-stream and therefore no wearing down of parts.
• Video capture equipment is fairly cheap.
• The instrument can operate continuously.
• Results are consistent for the same scene and can be directly compared with results from other locations.
• There are no mechanical parts which reduces maintenance due to wear and tear.

Results have been obtained with an average true detection rate of 69% and a further close miss rate of 14%, with very few false alarms (under various conditions of appearance and lighting). The overall result is that the measured distribution closely matches the true value for each of the test images.

BibTeX Bibliography

PhD Thesis

• Preamble (61kB)
• Chapter 1 Introduction (77kB)
• Chapter 2 Problem Statement (366kB)
• Chapter 3 Background Theory (437kB)
• Chapter 4 Intrument Structure (40kB)
• Chapter 5 Image Formation (476kB)
• Chapter 6 Image Preprocessing (511kB)
• Chapter 7 Macroscopic Blob Segmentation (268kB)
• Chapter 8 Blob Classification for Rock Detection (172kB)
• Chapter 9 Consolidation of Multiscalar Results to Separate Rocks (277kB)
• Chapter 10 Results (2263kB)
• Chapter 11 Conclusions (26kB)
• Postamble (65kB)

Papers

• Rock Detection Using Neural Networks
  • Abstract
  • Postscript paper
• Rock Recognition Using Feature Classification
  • Abstract
  • Postscript paper
• Multiscalar Rock Recognition using Active Vision
  • Abstract
  • Postscript paper

Bottletop Seal Inspection

Researcher

Trevor Bartleet assisted by Robert Crida (System software and algorithm development), Peter Moon (Initial algorithm development and software industrialisation), Andre le Roux (Early software development) and Anthony Kanfer (Early lighting design).

In collaboration with

Electronic Development House (EDH), Stellenbosch

Sponsor

FRD Manufacturing Grant

Dates

1993 to 1995

Synopsis

Plastic bottle tops or closures are produced for the bottling industry. A large proportion of the factory's production is exported to Europe and the Middle East during winter when the local demand is low. A number of defects can occur during the prodution of the closures. These defects can allow gas to escape from the bottle. In order for the closures to be acceptable to the bottling companies, they need to be of high quality with a low defect count (a fraction of one percent). This involves the individual inspection of every closure.

Human inspection of the closures has proved to be unreliable and subjective, given the difficult working conditions (closures are manufactured at a rate of 20 per second by each machine). A machine vision system is capable of producing reliable and repeatable results in an environment which is unpleasant to humans. System development was divided into three sections:

1. Design and selection of hardware including: camera, lighting, digital signal processing board, the host computer and transducers which interact with the environment.
2. The design of image processing algorithms to detect defects in the closures. The method used involves tracking the boundary of the liner.
3. Software for the host computer which manages the overall system including: user interface, communication with the DSP, the interactions with the environment and the timing and syncronisation of the system.

Results have been obtained where 99.9% of the good closures are accepted and more than 90% of the defective closures are removed. These results have been obtained using a system installed on a production line.

Papers

• Real Time Machine Vision Inspection of Plastic Bottle Closures
  • Abstract
  • Postscript paper

Fruit Quality Inspection

Researcher

Karen Henry

Sponsor

FRD Manufacturing Grant

Dates

1994 to 1995

Synopsis

South Africa has a very successful and lucrative fruit industry. Some of the fruit grown is sulphur dried and sold on the dried fruit market. The quality control of both the fresh and dried fruit is very important. The grading of fruit can be done either manually or automatically. An automated process will improve the speed of grading thereby enabling every single fruit to be tested, instead of a select few, as occurs in manual grading. Image processing on colour images of each fruit would achieve this.

Several such fruit sorting machines are available on the international market, at a cost of over R0,5 million. These machines are very successful in sorting certain types of dried fruit and particularly unsuccessful at others. A system was needed that:

1. ensured acceptable processing speed
2. catered for problem areas that currently caused dried fruit sorting, by image processing, to fail. These are:
   1. darker core areas of dried peaches cause peach to be rejected as rotten
   2. blemish areas, and grading of fruit by blemish sizes.

Papers

• Colour Image Analysis for Classification of Dried Fruit
  • Abstract
  • Postscript paper

Flotation Cell Surface Froth Analysis

Researcher

Jia Liu

Dates

1992 to 1995

Synopsis

In the mining industry, the froth structures on the surfaces of flotation tanks provide substantial visual information regarding the grade and recovery of the mineral being extracted. Consequently, the froth bubble size and shape distributions prove descriptive of the froth structure. The ability to accurately and reliably measure froth structure size and shape distributions using computer vision is of considerable interest to the industry and such a technique could have far reaching consequences in the optimisation of the control of industrial flotation cells.

Past research work in the field has resulted in a number of different techniques and algorithms for the purpose of segmentation. However, these algorithms are often iterative and therefore quite slow.The technique being developed is non-iterative and therefore one with industrial real-time processing implications.

A statistical comparison of the expected and observed distributions of bubble size, circularity and eccentricity revealed that the machine vision system does not mimic the human visual process exactly, but does however still achieve acurate segmentation which is favourable with regard to the expected results. In the final analysis, the computed parameters of bubble sizes and eccentricity can be used as robust and reliable measurements to describe the froth surface and bubble shapes. These parameters may thus be used to describe quantitatively the mineral grade and recovery values and be used in the optimsation of control of industrial flotation cells.

For further information please contact Professor Gerhard de Jager at