This invention relates to a method for measuring the volume of material on moving conveyors, and the apparatus therefor. Although the invention will be described with reference to the measurement of coal or overburden in an open-cut mining situation, where severe difficulties have to be overcome (for instance, where conveyor belts are up to 2 metres wide, where 24 hour operation is required, operation is in harsh open air conditions, and where burden surface is extremely variable in shape) , the present invention may be applied to volume-measurement on any bulk conveyor system, and is easy to apply in other cases where operating conditions are less harsh- Mined ores and other materials such as coal and overburden are transported out of open-cut mines by conveyor systems. Although the mass transport rate may be estimated using radiation penetration measurements, accurate volume transport rate determination has not been practicable.

The present invention is directed at providing an effective and practical means for the measurement of conveyor material volume throughput. According to the present invention there is provided apparatus for measuring the volume of material passing along a conveyor, comprising a light source to project a beam of light across said conveyor such that the leading edge of the beam of light impinges on material on the conveyor at a first angle, means to receive an image of at least the leading edge of the beam of light at a second angle, said second angle being displaced from said first angle and means to capture, and process, the image at a rate necessary to provide calculation of a required volume parameter.

The present invention also provides a method for measuring the volume of material passing along a conveyor comprising providing a beam of light from a light source across a conveyor such that the leading edge of the beam

SUSSTITUTΠ S _Ξ Γ

of light impinges on material on the conveyor at a first angle, receiving an image in an image receiving means of at least the leading edge of the beam of light at a second angle, said second angle being displaced from said first angle and capturing and processing the image at a rate necessary to calculate a required volume parameter.

The light source of the present invention is preferably a strobe light source. The strobe light source may be mounted above the conveyor belt in such a way as to produce, when triggered, a pulse illuminated zone on the conveyor, preferably in the form of a rectangular beam of light. Preferably the broad beam of light extends from side to side across the conveyor to cover the width of material being conveyed although if the accuracy required is less the beam may extend over the central portion of tHe conveyor only. Preferably the beam extends at about right angles to the direction of the conveyor belt travel. However, in some cases it may be preferred for the beam to extend diagonally across the conveyor belt for particular resolution of irregularly placed materials on the conveyor. Also the light source and/or the image recovering means may be located directly above the conveyor belt or displaced to one or other side to suit the expected disposition and size of the materials. The light source is preferably such as to provide a distinctly sharp leading edge across the material on the belt, and illumination of the material to one side of this leading, edge. In a preferred form, the light source is a powerful strobed xenon flash discharge light source, synchronised with the rest of the system in order to capture sharp, unblurred images.

The image receiving means is preferably a solid state camera utilising a charged coupled device (CCD) , or a similar target. Preferably the camera used is such as to inherently integrate the light received by its target, and provide a stable image without appreciable distortion or drift. The resulting image is captured and processed using an image-processing system. This procedure is repeated preferably several times per second so that

-s -> i_>Zt t i _ -A- f

~ _. on a movin ^g measurements can _b e de at fluent

_• m _{mr th} at' ^" t _h e wor _k ing zone is ^p rotecte ^d dayli _gh t, providing that ^t ne ii _rtV.t - _{«nι h} as by a simple open ^•» - ^.«• = ^« - ^• . from direct sunlight such By _operate d in a _i n a preferre _d form, the .system __ _

_« „ _rh as in a shed, or merely un ^d er a darkened area, such as ⁱ n - _ass ociate _d with _j cover _, in this way, the slight problems assoc ⁱ ate use in strong light are avoide _necha nical

It should be appreciated that ^t ne shutter may be omitted, and hooded enclosures may be use ^d

^t ne

^{thβ bel} ";: s ^' f , the primary processe ^d ima ^g e i _ _. effect^ a _{.5 d} e _g ree ._.nt section throu ^g h con ^v eyor and

I TE SK££

W

_{39 ma} y allow for greater flexi ^b in ^y

SUBSTiTUTE

tablet, video recorder, and a time-base corrector.

Separate instrumentation measures the instantaneous belt speed, and this may also be fed into the computer.

The signal from the image receiving means, preferably a standard interlaced video signal, is fed to the computer system, which may be fitted with commercially available image processing hardware, allowing images to be digitised, processed, and displayed. The system described above provides a resolution of approximately 512 x 512 pixels, which provides ample resolution for any conceived conveyor measurement system.

Images are preferably captured and processed at a rate necessary to provide accurate calculation of the volume rate and total running volume, and will depend on the speed of the belt and the expected variation in the material volume.

As each image is captured it is processed and the cross-sectional area of the material is calculated by measuring the area between the unloaded belt reference profile and the new edge profile on the current image. These calculations take into account lens distortion, the vertical scaling required due to the camera's angle of view, and the pixel/real world measurements' ratio which is calculated initially when the unloaded belt reference is taken.

The image is processed by searching across the image from the dark area until the light edge threshold is found, and recording the co-ordinates of that point. This is done for a number of lines across the belt image (approximately 100 has been found satisfactory) in order to build up a profile of the burden. This profile is compared with the unloaded belt reference profile to calculate the area of the burden for that image sample, taking into account known distortions. In some applications where the image signals are not sufficiently clear, due to signal noise for example, initial image enhancement of the image is required by using enhancement techniques such as noise filtering or edge enhancement. These can be carried out in the hardware on the image

f SUBSTITUTE SblΣZT

processing boards, or may be done in software.

Knowing the belt speed, and the time elapsed since the last image sample, the distance the belt has travelled and the volume of burden transported are then calculated. The calculation assumes that the cross sectional area is constant between samples; any error introduced by this assumption depends on the belt speed, and the sampling rate.

A dynamic visual computer generated display of the conveyor material preferably is then displayed on the high resolution monitor, together with a graphical presentation of the instantaneous volume rate and total burden volume transported.

In situations where no graphical output display and. - minimal output information is required, an alternative method of processing the video image from the camera may be employed. For example, where the only output required is the instantaneous or averaged cross section of the burden, then the video signal may be processed directly by analogue/digital circuitry and the output fed directly to a digital display unit. Such a system does not require the use of an imaging computer, and avoids one of the major costs of the measurement system. _______

In a preferred form this alternative method can be operated as follows. The video signal from the camera may be fed directly to specially designed analogue/digital circuitry which effectively measures the time from start of the line to the occurrence of the light -edge on each of the video image lines. The circuitry sums these measurements, and subtracts from the result, a previously measured reference measurement made on an empty belt.

After conversion the result is a measurement of the burden cross sectional area, which is then directly displayed on a simple digital display unit. Where multiple or complex outputs are required, which are difficult to produce from discrete circuitry, a minimal low cost specially designed microprocessor board may be used to do some of the processing. In order to assist in arriving at an understanding

SUeST ^. TUTE SH££T

of the present invention, several preferred embodiments are illustrated in the attached drawings. However, it should be understood that the f llowing description is illustrative only and should not be taken in any way as a restriction on the generality of the invention as described above.

In the drawings:

Figure 1 is a schematic diagram of the present invention; Figure 2 is a schematic diagram of a preferred Image Processing system for use with the present invention;

Figure 3 illustrates possible images as processed by the apparatus of the present invention, and utilised. • for—calculations; and

Figure 4 illustrates the masking effect created by non-uniform material when using the apparatus of the present invention.

Figure 1 shows the general arrangement of the principal parts of the apparatus of a preferred embodiment of the present invention. Illustrated is a conveyor belt 10 having thereon material or burden, 12. A strobe light source 14 provides a broad beam of light 16 on the surfacje- of the material 12. A camera 18 is arranged to receive the image of at least the leading edge 20 of the broad beam of light 16, to transfer the image to the processing means 22.

The strobe light source 14 is preferably based on a Xenon flash tube with a fused quartz envelope, which is about 300 mm long by about 10 mm diameter. It is rated for an average discharge power of about 500 watts with forced air cooling, and has an expected life of at least 50 million flashes at about 40 joules per flash. Satisfactory images are obtained at about 16 joules per flash and therefore a life of at least six months per tube in continuous operation at 24 hours per day and 5 flashes per second is expected.

The energy for the flash is stored in high grade, high peak current rated capacitors whic _h are

SUESTITUTΠ SI-^Γ

non-dissipatively charged from 240 volt mains. The flash is triggered by a lower power short duration firing pulse derived from the shutter and strobe- synchroniser, under control from the processing means 22. The shaped illumination pattern is formed by a full length slit, approximately 8mm wide located in front of the tube, and a one metre wide single sided gate at a distance of about 600 mm from the slit.

Preferably, a high resolution camera 18 is used which includes a power supply/electronics unit providing

AGC, Gamma and Aperture Correction. This unit has the advantage of providing a number of synchronising waveforms which are useful in connection with the general circuits designed to synchronise the camera, shutter, and light ^' source functions. The camera preferably has a standard

CCD target of 2/3 inches diagonal with a matrix of 550(H) by 582(V). The supply/electronics unit may be modified to provide wide range manual gain control.

The processing means 22 captures and processes the image using specialist operating software, and provides a display 24 on a display monitor 26. Connected to the processing means 22 is a terminal 28 to input data and information, together with a belt speed sensor 30, and—a' strobe trigger 32. Belt speed sensor 30 provides the processing means 22 with the speed of the conveyor belt in order to assist in calculations. The strobe trigger 32 is coordinated with a phase locked loop control 34 attached to the camera 18, so that the processing means can coordinate the triggering of the strobe light to emit the beam of light at the same time as a mechanical shutter provided in the camera is open.

The operating software preferably consists of two programs. The first program calculates certain parameters and stores them on disc. The main operating program retrieves this information when first run and uses this data to search for the light edge and to calculate the necessary areas and volumes.

The first program is run prior to the start of the main operating program and need only be run again if some

SUDSTin;T2 SH£ET

change has occurred to the relative positions of the camera, light source and conveyor belt. This program calculates, from a number of empty belt images, the following parameters: . the empty belt reference position - as a string of co-ordinates; the vertical and horizontal pixel calibrations (mm/pixel); the position of . the belt images, and hence the co-ordinates of a search frame.

Figure 3 illustrates a typical empty belt image 62, with the search frame shown as the rectangular border. The sides of the search frame running across the belt are set to a position which ensures that the belt illuminated. ^• light edge always occurs within these lines. The other limits are calculated after searching across the full width of the frame to determine the edges of the belt. The pixel calibrations are calculated from the known distances between the edges of the belt and from the bottom of the belt to the mean height of the two edges.

In a preferred form, the main operating program works as follows. After completing the processing of each image a trigger pulse is sent to the strobe light at _^ tlie next appropriate point in the repetitive television camera signal readout sequence and the illuminated image is recorded on the camera target. The image is then read out from the camera over approximately the next 40 milliseconds in broadcast television format, and recorded into a digital image storage buffer for subsequent analysis.

The computer then commences searching the stored image for the "first illuminated edge", scanning from below the empty belt position towards the maximum expected burden height in the previously defined search area. Figure 3 also depicts the image obtained with a loaded belt 64.

The search is done at a large number of positions across the belt width (approximately 100) and the software calculates the height of the burden above the

SUBSTITUTE SHi___T

corresponding empty belt'"position _a t each lateral step. It then calculates the corrected area for each of these incremental sections making use of lateral and vert ⁱ ca calibration factors which were obtained dur ⁱ ng the ⁱ n ⁱ t ⁱ al calibration program. On completion of the search steps these areas are sui-ed to produce a total cross sectional area for that measurement point.

The burden volume between the current and the previous measurement point is then calculated by Multiplying the burden cross sectional area just computed _by the distance of the belt travel between the two measurement points. In this preferred form a constant belt speed was assumed and the distance computed from the belt speed and the elapsed time between the two. measurements. _M-ι4 _ .

Since burden section of adjacent measurement po ⁱ nts is often variable, an integrated running average of the volume throughput rate is derived from the last ten processed images for display purposes.

The total throughput volume for the conveyor is recorded by summing the burden volumes which have been calculated between each measurement point.

On completion of the analytical scan of the ft ⁱ U ^r - belt width image, the computer displays the averaged volumetric rate and the total throughput volume. It then plots the same information on a graph and generates a graphical display of the burden shape on the belt (see Figure 1, numerals 24 and 26).

The strobe trigger 32, phase locked loop control 34, and the shutter on the camera 18 are optional features which may be utilised when operating in extremely br ⁱ ght conditions, or with fast moving belts. They may also be used in other conditions if desired. The strobe l ⁱ ght source may also be enclosed in a rapidly detachable sub-assembly which will ensure precise location and easy field replacement. Such an assembly is preferably totally sealed, and will passively dissipate waste heat from the outside of the unit.

The shutter preferably consists of a disc rotat ⁱ ng

SUBSTITUTE SKSET

in the space between the lens and the camera imaging target, driven by a d.c. motor at a speed of about 1500

R.P.M., i.e. about 25 revolutions per second, corresponding to the television camera frame rate. The camera/shutter unit is also preferably fully sealed, and made tough enough to tolerate frequent bursts of high pressure wash down water. It may also be provided with an automatic window cleaning system, to be operated on a regular time schedule. The strobe light source 14 is here illustrated projecting a beam of light at an angle of about 45 to the level of the conveyor belt, and the camera receives the image also at an angle of about 45 to the level of the conveyor belt. _ A preferred processing means 22 is further illustrated in Figure 2. Figure 2 illustrates a micro VAX2 based image processing system having image processing boards 40 to 50, from the IP512 family of Imaging Technology Inc., which include: an AP512 image digitiser (40); an ALU512 arithmetic and logic unit (42) incorporating pipelined multiplication, logical operations such as addition, subtraction, etc., and barrel shifter; three FB512 frame buffers (44,46,48); and an HF512 histogram and feature extraction. unit (50), where "feature extraction" consists of generating a list of the indices of pixels of specified intensities. The IP512 boards process 512 x 512 x 8 bit black and white images from composite video sources at full frame rate.

Auxilliary equipment illustrated in Fig. 2 includes 31cm and 48cm color monitors 52, two black and white cameras 54, a video cassette recorder 56 with a good single frame advance, a time base corrector 58, and a digitising tablet 60.

Figure 4 illustrates a section 70 of the material 12 illuminated by the broad beam of light 16. In this section the leading edge of the illuminated material is obscured from the view of the camera 18. This has the effect of making the detected pseudo edge 72 different to the actual leading edge 20. If the light source projected

SUUSBSSrΪ .v YE sμjsc-

a thin stripe of light at the leading edge, instead of a broad beam of light as illustrated, no measurement could be made. The use of a broad beam, of light overcomes this problem.

5 Therefore, the apparatus and method of the present invention employs an active lighting technique coupled with non-contact image capture, processing, and display output equipment, in order to measure the volume of coal and overburden, for example, on moving conveyors (both

10 fixed and on dredger conveyors) with more accuracy and reliability than existing methods.

15

20

^' 25

30

35

39

SUBSTITUTESHEET