CN103364315B - A kind of sintering solid fuel granularity online test method and pick-up unit - Google Patents

Abstract

The invention provides a kind of sintering solid fuel granularity online test method and pick-up unit, detection method comprises: solid-fuelled smooth pressing, the image data acquiring of granularity, the image data extraction of granularity, the pre-processing image data of granularity, the Iamge Segmentation process of granularity, the image characteristics extraction of granularity, the characteristics of image Classified statistics of granularity.Pick-up unit comprises: with the fuel belt of side carrying roller choosing one section and arrange the straight carrying roller of a row below fuel belt, at the intersection of side carrying roller and straight carrying roller, one piece of scraper plate is set, one slicking-in roller is set after scraper plate, above fuel belt, arrange a light source impinges upon on fuel planar to be measured, arranges an image capture device above fuel planar to be measured.The problems such as the present invention carries out image procossing, feature extraction and analytical calculation by computing machine, detects solid fuel size-grade distribution density, and solve manual detection labour intensity large, detection time is long, real-time property difference.

Description

A kind of sintering solid fuel granularity online test method and pick-up unit

Technical field

The present invention relates to solid fuel granularity Detection technology, especially a kind of sintering solid fuel granularity online test method and pick-up unit.

Background technology

In sintering process, the fluctuation of solid fuel granularity directly causes its surface area to change, make the change of the wetting of carbon granules self and ignition process, also sinter mixture pelletizing effect can be affected, bring impact to be that the parameter such as density, voidage bulk density of compound changes on the physico-chemical property of compound, cause that whole mixture ventilation is uneven, skewness walked crosswise by sintering combustion band.In addition, solid fuel change of granularity can cause that solid fuel ignition dynamics in sinter mixture is corresponding with thermodynamic condition to change, thus affect the product of burning rate and charcoal burning, above-mentioned factor impacts SINTERING PRODUCTION all to a great extent, and sinter quality and output are reduced.For solving solid fuel fluctuation problem, first need solid fuel granularity on-line checkingi, by on-line checkingi solid fuel granule size, adjustment disintegrating machine (roll gap), control solid-fuelled particle size after cracking, make solid fuel granularity meet manufacturing technique requirent.Traditional solid fuel detection method manually samples from the fuel feed belt after fragmentation, and off-line application combination sieve carries out sieving, weigh, add up and obtain solid-fuelled granularity data.This method is whole manual operations, and labour intensity is large, and detection time is long, and data are poor in real time, and cannot realize informationization, Aulomatizeted Detect controlling functions.

Summary of the invention

The object of the present invention is to provide a kind of sintering solid fuel granularity online test method and pick-up unit, adopt the solid fuel particle size detection method of computer machine vision and graphical analysis, by being arranged on material pressuring flat device above delivery solid fuel belt feeder and image capture device carries out Real-time Collection to solid fuel granularity image, image procossing is carried out by computing machine, feature extraction and analytical calculation, detect solid fuel size-grade distribution density, solve the manually-operated labour intensity of solid fuel in SINTERING PRODUCTION technique large, detection time is long, the problems such as real-time property difference, for realizing the robotization that solid fuel fragmentation controls, informationization provides detection data.

The object of the present invention is achieved like this, and a kind of sintering solid fuel granularity online test method comprises the following steps:

(1) solid-fuelled smooth pressing, applies the material scraper plate and slicking-in roller that are arranged on above solid fuel feed belt, is put down by solid fuel flat seamless, detect in order to image.

(2) solid fuel granularity image data acquisition, the solid-fuelled image of pressing is flattened by the image capture device on-line checkingi be arranged on above solid fuel feed belt, its sampling period arranges and matches with solid fuel feed belt travelling speed, avoids solid fuel image fault or its image information is lost.Adopt single bright field illumination, without the need to details in a play not acted out on stage, but told through dialogues, like this while simplified apparatus, the more important thing is and the contrast of solid fuel granularity image is enhanced, be more conducive to the accurate of solid fuel granularity Detection.

(3) solid fuel granularity image data are extracted, by the sampling period of computing equipment by image capture device, the solid fuel surface realtime graphic collected is extracted, then one is created to each width target image and independently analyze thread, to carry out further analyzing and processing.

(4) solid fuel granularity image data prediction, adopts medium filtering and gray-level histogram equalization method to carry out noise reduction process and image optimization to image, avoids the noise in solid fuel image to impact solid-fuelled Detection results.

(5) solid fuel granularity image dividing processing, adopts the method for adaptive threshold fuzziness to carry out Iamge Segmentation according to the feature of solid fuel image, determines solid fuel granularity boundary image.The determination of solid fuel granularity boundary image utilizes threshold value to split by gray level image, and threshold value have employed the adaptive threshold relevant to solid fuel granularity image average gray, and the coefficient of adaptive threshold is obtained by a large amount of on-site solids fuel granularity view data test, effectively can determine the initial starting point on solid fuel granularity border.

(6) solid fuel granularity image feature extraction, extracts solid-fuelled two features in image, is the area of the approximate length and width when particle of solid fuel particle respectively.

(7) solid fuel granularity image characteristic statistics and analysis, according to SINTERING PRODUCTION technique initialization solid fuel granule size rank level, with solid-fuelled two eigenwerts for dividing condition, carry out classification cycle statistical computation, calculate quantity and the number percent thereof of each level solid fuel granule size, in this, as solid fuel granularity Detection analysis result.

A kind of pick-up unit for realizing above-mentioned sintering solid fuel granularity online test method, with the fuel belt of side carrying roller choosing one section and arrange the straight carrying roller of a row below fuel belt, at the intersection of side carrying roller and straight carrying roller, one piece of scraper plate is set, one slicking-in roller is set after scraper plate, above fuel belt, arrange a light source impinges upon on fuel planar to be measured, above fuel planar to be measured, arrange an image capture device, described image capture device is camera or high-resolution video camera.

The present invention adopts the solid fuel particle size detection method of computer machine vision and graphical analysis, image procossing, feature extraction and analytical calculation is carried out by computing machine, detect solid fuel size-grade distribution density, solve manual detection labour intensity large, detection time is long, the problems such as real-time property difference, achieve solid fuel granularity on-line checkingi, for robotization, the informationization realizing solid fuel fragmentation control provides detection data.The principle of the invention is simple and clear, and adapt to site environment, reliable operation, equipment cost is low, is easy to realize.

Accompanying drawing explanation

Fig. 1 is present device pie graph;

Fig. 2 is target particles feature schematic diagram.

In figure: 1, belt roller; 2, feed belt; 3, coal dust; 4, scraper plate; 5, slicking-in roller; 6, belt flat carrier roller; 7, video camera.

Embodiment

As shown in Figure 1, a kind of sintering solid fuel granularity on-line measuring device, with the fuel belt 2 of side carrying roller 1 choosing one section and arrange the straight carrying roller 6 of a row below fuel belt 2, at side carrying roller 1 and the intersection of straight carrying roller 6, one piece of scraper plate 4 is set, one slicking-in roller 5 is set after scraper plate 4, above fuel belt 2, arrange a light source impinges upon on fuel planar to be measured, above fuel planar to be measured, arrange an image capture device 7, described image capture device is camera or high-resolution video camera.

A kind of sintering solid fuel granularity online test method of the present invention comprises the following steps:

(1) solid-fuelled smooth pressing, when solid fuel is by system equipment, solid fuel is struck off also compacting, pressing by the material scraper plate that application is arranged on above solid fuel feed belt and slicking-in roller, detects in order to image.

(2) solid fuel granularity image data acquisition, the solid-fuelled image of pressing is flattened by the image capture device on-line checkingi be arranged on above solid fuel feed belt, its sampling period arranges and matches with solid fuel feed belt travelling speed, avoids solid fuel image fault or its image information is lost.Adopt single bright field illumination, without the need to details in a play not acted out on stage, but told through dialogues, like this while simplified apparatus, the more important thing is and the contrast of solid fuel granularity image is enhanced, be more conducive to the accurate of solid fuel granularity Detection.

(3) solid fuel granularity image data are extracted, by the sampling period of computing equipment by image capture device, the solid fuel surface realtime graphic collected is extracted, then one is created to each width target image and independently analyze thread, to carry out further analyzing and processing.

(4) solid fuel granularity image data prediction, in order to retain the clear of image outline, first carries out noise reduction process based on medium filtering to the coal dust image collected.For the coal dust image collected substantially pixel grey scale be all darker, be therefore then to noise reduction after image carry out gray-level histogram equalization process, increase the dynamic range of gradation of image, strengthen contrast.

(5) solid fuel granularity image dividing processing, the present invention adopts the method for following adaptive threshold fuzziness to carry out Iamge Segmentation according to the feature of coal dust image, determines pulverized coal particle image, and wherein the determination algorithm of segmentation threshold is as follows:

1. a threshold value T is set _k, image is divided into two groups of R1 and R2

R ₁＝{f(x，y)|f(x，y)≥T _k}

R ₂＝{f(x，y)|0＜f(x，y)＜T _k}

Wherein, the gray-scale value put for (x, y) on image of f (x, y).

2. μ is calculated _k(threshold distance coefficient)

${\mathrm{\μ}}_k=|\frac{\underset{(x,y)\∈R_1}{\mathrm{\Σ}}f(x,y)/N\left(R_1\right)+\underset{(x,y)\∈R_2}{\mathrm{\Σ}}f(x,y)/N\left(R_2\right)}{2}-T_k|$

Wherein N (R ₁), N (R ₂) be respectively the average gray in R1 region and R2 region.

3. final segmentation threshold AAK is calculated

AAK is for work as μ _kget T during minimum value _kvalue.

(6) solid fuel granularity image feature extraction, the present invention extracts two features of pulverized coal particle in image, be respectively particle be similar to length and width when particle-surface amass.

The computing method that particle is similar to length breadth ratio RatioL/W are as follows:

$\mathrm{RatioL}/W=\frac{\mathrm{LineAB}}{\mathrm{LineCD}}$

Particle approximate length LineAB be in Fig. 2 target particles image extraneous rectangle diagonal line through the longer length of one of two line segments of target particles image section, particle approximate width LineCD be in Fig. 2 the extraneous rectangle diagonal line of target particles image through the shorter length of one of two line segments of target particles image section.

$\mathrm{LineAB}=\sqrt{{\left({\left|\mathrm{AB}\right|}_x\right)}^2+{\left({\left|\mathrm{AB}\right|}_y\right)}^2}$

$\mathrm{LineCD}=\sqrt{{\left({\left|\mathrm{CD}\right|}_x\right)}^2+{\left({\left|\mathrm{CD}\right|}_y\right)}^2}$

Wherein | AB| _xfor AB WAWQ square to spaced image vegetarian refreshments number, | AB| _yfor AB two selects the spaced image vegetarian refreshments number of vertical direction.Particle area is the number of the pixel of dash area in Fig. 2.

(7) solid fuel granularity image characteristic statistics and analysis, according to SINTERING PRODUCTION technique initialization solid fuel granule size rank level, with solid-fuelled two eigenwerts for dividing condition, carry out classification cycle statistical computation, calculate quantity and the number percent thereof of each level solid fuel granule size, in this, as solid fuel granularity Detection analysis result.

First all particles are once sieved, approximate length breadth ratio (RatioL/W) as fruit granule is less than 10 and particle area is greater than smallest particles area (MinArea), then this particle can carry out next step statistics, otherwise then cannot carry out next step statistics.Wherein MinArea choosing method is:

First selected value meets following condition

And the final selected value of MinArea is all maximal values met in the selected value of above formula.

Then again sieve remaining particle target according to particle approximate length, all target particles are sieved into 6 groups by this screening, and the boundary condition of 6 groups is amplification situation according to real image and determines, and concrete method for sieving is as follows:

Wherein: BG1 is the display length of 1mm physical size on image, BG2 is the display length of 2mm physical size on image, BG3 is the display length of 3.15mm physical size on image, BG4 is the display length of 5mm physical size on image, and BG5 is the display length of 6.3mm physical size on image.

Add up solid fuel size-grade distribution density at different levels after screening respectively, distribution density statistical method is as follows:

System carries out comprehensive statistics to continuous 5 particle plane pictures, data are exported as final solid fuel granularity Detection using the distribution density average of each level solid fuel granularity, carry out online production Operating Guideline, and then adjustment relevant device parameter, control the quality of solid fuel granularity and sintering deposit.

Claims (2)

1. a sintering solid fuel granularity online test method, is characterized in that comprising the following steps:

(1) solid-fuelled smooth pressing, utilization is arranged on the straight carrying roller below material scraper plate, slicking-in roller and the belt above solid fuel feed belt, is put down by solid fuel flat seamless, detects in order to image;

(2) solid fuel granularity image data acquisition, the solid-fuelled image of pressing is flattened by the image capture device on-line checkingi be arranged on above solid fuel feed belt, its sampling period arranges and matches with solid fuel feed belt travelling speed, and adopts single bright field illumination;

(3) solid fuel granularity image data are extracted, by the sampling period of computing equipment by image capture device, the solid fuel surface realtime graphic collected is extracted, then one is created to each width target image and independently analyze thread, to carry out further analyzing and processing;

(4) solid fuel granularity image data prediction, adopts medium filtering and gray-level histogram equalization method to carry out noise reduction process and image optimization to image;

(5) solid fuel granularity image dividing processing, the method of adaptive threshold fuzziness is adopted to carry out Iamge Segmentation according to the feature of solid fuel image, the coefficient of adaptive threshold is obtained by a large amount of on-site solids fuel granularity view data test, effectively can determine the initial starting point on solid fuel granularity border, concrete grammar is:

1. a threshold value T is set _k, image is divided into two groups of R1 and R2

R ₁＝{f(x,y)|f(x,y)≥T _k}

R ₂＝{f(x,y)|0<f(x,y)<T _k}

Wherein, the gray-scale value put for (x, y) on image of f (x, y);

2. μ is calculated _k

${\mathrm{\&mu;}}_k=|\frac{\underset{\left(x,y\right)\&Element;R_1}{\&Sigma;}f\left(x,y\right)/N\left(R_1\right)+\underset{\left(x,y\right)\&Element;R_2}{\&Sigma;}f\left(x,y\right)/N\left(R_2\right)}{2}-T_k|$

Wherein N (R ₁), N (R ₂) be respectively the pixel number in R1 region and R2 region;

3. final segmentation threshold AAK is calculated

AAK is for work as μ _kget T during minimum value _kvalue;

(6) solid fuel granularity image feature extraction, extracts solid-fuelled two features in image, is the area of the approximate length and width when particle of solid fuel particle respectively;

The computing method that solid fuel particle is similar to length breadth ratio RatioL/W are as follows:

$RatioL/W=\frac{LineAB}{LineCD}$

LineAB is target particles image extraneous rectangle diagonal line through the longer length of one of two line segments of target particles image section, and LineCD is the extraneous rectangle diagonal line of target particles image through the shorter length of one of two line segments of target particles image section;

$LineAB=\sqrt{{\left(\right|AB{|}_x)}^2+{\left(\right|AB{|}_y)}^2}$

$LineCD=\sqrt{{\left(\right|CD{|}_x)}^2+{\left(\right|CD{|}_y)}^2}$

Wherein | AB| _xfor AB WAWQ square to spaced image vegetarian refreshments number, | AB| _yfor AB two selects the spaced image vegetarian refreshments number of vertical direction; | CD| _xfor CD WAWQ square to spaced image vegetarian refreshments number, | CD| _yfor CD two selects the spaced image vegetarian refreshments number of vertical direction; The pixel number that particle area comprises for each solid fuel particle on image;

(7) solid fuel granularity image characteristic statistics and analysis, according to SINTERING PRODUCTION technique initialization solid fuel granule size rank level, with solid-fuelled two eigenwerts for dividing condition, carry out classification cycle statistical computation, calculate quantity and the number percent thereof of each level solid fuel granule size, in this, as solid fuel granularity Detection analysis result;

Concrete grammar is as follows:

First once sieve all particles, the approximate length breadth ratio RatioL/W as fruit granule is less than 10 and particle area is greater than smallest particles area MinArea, then this particle carries out next step statistics, otherwise does not then carry out next step statistics; Wherein MinArea choosing method is:

First selected value meets following condition

And the final selected value of MinArea is all maximal values met in the selected value of above formula;

Then again sieve remaining particle target according to particle approximate length, add up solid fuel size-grade distribution density at different levels after screening respectively, distribution density statistical method is as follows:

2. one kind for realizing the pick-up unit of sintering solid fuel granularity online test method according to claim 1, it is characterized in that with the fuel belt (2) of side carrying roller (1) chooses one section and fuel belt (2) below the straight carrying roller of one row (6) is set, at side carrying roller (1) and the intersection of straight carrying roller (6), one piece of scraper plate (4) is set, one slicking-in roller (5) is set after scraper plate (4), arranging a light source in fuel belt (2) top impinges upon on fuel planar to be measured, an image capture device (7) is set above fuel planar to be measured, described image capture device is camera or high-resolution video camera.