CN101419151B - Industrial pure terephthalic acid particle size distribution estimation method based on microscopic image - Google Patents

Abstract

The invention discloses a new method for estimating the particle diameter distribution of industrial pure terephthalic acid based on a microscopic image. The method comprises the following main steps: (1) the magnification times of the image is indicated; (2) an image of the industrial pure terephthalic acid is acquired through a microscope and a CCD camera; (3) the number, the area and the largest particle diameter of industrial pure terephthalic acid particles in a single image are extracted; (4) the area of the industrial pure terephthalic acid particles is utilized to obtain the equivalent diameter and the equivalent size; (5) a plurality of images are operated to obtain the particle diameter distribution and the size distribution; and (6) results of the particle diameter distribution and the size distribution are fitted to obtain the real particle diameter distribution and the real size distribution. The invention provides a set of a system which has low cost and high measuring precision and is applied to the mensuration and control of the particle diameter distribution in the industrial pure terephthalic acid industry.

Description

A kind of industrial pure terephthalic acid particle size distribution estimation method based on micro-image

Technical field

The invention belongs to the industrial control technology field, the measuring method that is to use that relates to is measured grain diameter, particularly a kind of industrial pure terephthalic acid particle size distribution estimation method based on micro-image.

Background technology

Industrial pure terephthalic acid is important Organic Chemicals, is mainly used in the production polyester product.The production of the size of industrial pure terephthalic acid particle diameter and distribution situation and polyester has close getting in touch, and influences the speed of esterification on the one hand; Influence the viscosity and the stability of slurry on the other hand.So the manufacturing enterprise of industrial pure terephthalic acid and user measure the size distribution of industrial pure terephthalic acid product and control as important index.The most frequently used measuring method of industrial pure terephthalic acid size distribution mainly contains two kinds at present, comprises laser particle analyzer method and sieve method.

The principle of laser particle analyzer is based on the scattering or the diffraction of laser; The size of particle shows through the size of scattering angle; Granule is big to the scattering angle of laser; Oarse-grained scattering angle is little, through the measurement (stacks of different scattering particles) to particle angle scattered light intensity, uses inverse matrix to decompose the size distribution that the angle scattered light intensity can obtain sample again.See that from principle the grain diameter that the laser particle analyzer test obtains is equivalent to the sphere diameter of identical angle of diffraction, it is the function of particle cross-sectional area.But the cross-sectional area of irregular particle usually can be greater than the spheroid of equal volume, and therefore, the diameter of the irregular particle that laser particle analyzer is measured is bigger than the spheric grain of equal volume.

The advantage of laser particle analyzer is that the repeatability of test result is better, and measuring accuracy is higher.But laser particle analyzer can not obtain the true particle diameter of the outer particle of range.For example Mastersizer S laser particle analyzer range is at 0.05～900 μ m, this means in sample less than 0.05 μ m and greater than the granule content of 900 μ m more for a long time, measurement result has than mistake.The another one inferior position of laser particle analyzer is the price that it is expensive, so also do not popularize at home.Because the technology content of laser particle analyzer is very high, have only in the countries such as some days, English, moral, U.S.A the rich enterprise of technology to produce at present, home products also can't be compared with it aspect many with stable at technical parameter.

Except laser particle analyzer, the another kind of grain diameter measurement method of often using was a sieve method during industrial pure terephthalic acid was produced.This method is by the WS, but directly under the drying solid state, with the sieve in different apertures with other particle of different-grain diameter level directly screening open, the gained particle diameter is commonly called screening particle diameter or sedimentation diameter.Specific practice is by vibrating device, and quantity of sample through a series of standard sieves with different sieve diameters, is separated into several grades, weighs respectively again, tries to achieve the industrial pure terephthalic acid particles size distribution of representing with massfraction.The advantage of sieve method is that principle is simple, and is easy to operate, cheap, and test process is simple and clear, and uncertain factor is few.

But; Aspherical particle depends on the orientation of particle to a great extent through the situation of sieve, because the industrial pure terephthalic acid sample is easy to generate static in vibration processes, a part of sample can be attached on the metallic screen; Be difficult for through sieve aperture, thereby cause measurement data inaccurate.Another shortcoming of sieve method is to be the quality summation of each particle diameter grade industrial pure terephthalic acid particles through the measurement result that sieve method obtains, and can't obtain the definite parameter of each industrial pure terephthalic acid particles.

Summary of the invention

The objective of the invention is to overcome the deficiency of prior art, a kind of industrial pure terephthalic acid particle size distribution estimation method based on micro-image is provided.

New particle size distribution estimation method based on the industrial pure terephthalic acid micro-image comprises the steps:

1) stereomicroscope imaging system total magnification is set to 100 times, and wherein the enlargement factor of CCD camera is 25 times, and the enlargement factor of objective lens is 2 times, and regulating zoom and focusing knob enlargement factor is 2 times;

2) micrometer is placed on the objective table, regulate the rising-falling tone focus knob,, utilize computing machine to obtain the true length of each pixel representative in the micrometer image up to observing micrometer image fully clearly;

3) the industrial pure terephthalic acid powder is put into planchet, place vibrating device to handle after 1-2 minute and be transferred on the black backlight of objective table;

4) regulate the rising-falling tone focus knob up to fully clearly observing industrial pure terephthalic acid particles, preserve image to computing machine;

5) be gray level image with the industrial pure terephthalic acid image transitions, carry out binary conversion treatment and Filtering Processing, extract number, true area and the maximum particle diameter of industrial pure terephthalic acid particles in the image;

6) with the true area S of industrial pure terephthalic acid particles _Actual(i), i=1,2 ..., n converts the equivalent diameter and the volume of each industrial pure terephthalic acid particles into, and computing formula is following:

$S_{\mathrm{actual}}\left(i\right)=\frac{\mathrm{\π}}{4}{D_{\mathrm{actual}}}^2\left(i\right)\⇒D_{\mathrm{actual}}\left(i\right)=\sqrt{\frac{4}{\mathrm{\π}}\×S_{\mathrm{actual}}\left(i\right)}$

$V_{\mathrm{actual}}\left(i\right)=\frac{\mathrm{\π}}{6}{D_{\mathrm{actual}}}^3\left(i\right),$ i＝1，2，...，n

D wherein _Actual(i) and V _Actual(i) be respectively the equivalent diameter and the volume of i industrial pure terephthalic acid particles;

7) repeating step 3) to step 6), obtain the equivalent diameter and the volume of several industrial pure terephthalic acid images;

8) utilize the equivalent diameter of industrial pure terephthalic acid to obtain the size distribution and the volume distributed median of industrial pure terephthalic acid;

9) utilize standardized normal distribution, skewness distribution and lognormal distribution that industrial pure terephthalic acid size distribution and volume distributed median result are carried out match, choose true size distribution and the true volume distribution of the minimum distribution of error sum of squares respectively as industrial pure terephthalic acid.

Said step 2):

(1) micrometer is placed on the white backlight plate, regulates the rising-falling tone focus knob, find after the outer circle of scale mark, seek scale mark and scale mark is moved to visual field central authorities; Regulate the rising-falling tone focus knob up to fully clearly observing each scale mark, preserve image to computing machine;

(2) be gray level image with the image transitions that obtains, selecting the reference gray level threshold value is that 50 pairs of gray level images carry out binary conversion treatment, wherein gray-scale value greater than this threshold value be the background area, otherwise be the target area;

(3) utilize the ordfilt2 function of Matlab that image is carried out Filtering Processing, parameter is chosen as the matrix of 6 row, 1 row;

(4) filtered bianry image is carried out line scanning and column scan, find out each target area, wherein the label of scale mark is K1; K2 ..., Kn; The coordinate figure of each Regional Representative's frontier point is kept in the corresponding matrix, and K1 number the centre coordinate with the Kn scale mark obtains through following formula:

$\stackrel{\‾}{X\left(p\right)}=\frac{1}{m}\underset{i=1}{\overset{m}{\mathrm{\Σ}}}X\_\mathrm{coordinate}(p,i),$ $\stackrel{\‾}{Y\left(p\right)}=\frac{1}{m}\underset{i=1}{\overset{m}{\mathrm{\Σ}}}Y\_\mathrm{coordinate}(p,i)$ ，p＝K1，Kn

X (p) wherein, Y (p) is the coordinate of pixel in the image, and to be each regional coordinate preserved the horizontal ordinate of scale mark to number X_coordinate to m, and Y_coordinate has preserved the horizontal ordinate of scale mark;

(5) pixel count between K1 scale mark and the Kn scale mark coordinate center obtains through following formula:

$\mathrm{Pixels}=\sqrt{{(\stackrel{\‾}{X\left(K1\right)}-\stackrel{\‾}{X\left(\mathrm{Kn}\right)})}^2+{(\stackrel{\‾}{Y\left(K1\right)}-\stackrel{\‾}{Y\left(\mathrm{Kn}\right)})}^2},$

The true length of each pixel representative obtains through following formula:

$1\mathrm{pixel}=\frac{1}{\mathrm{Pixels}}\mathrm{mm}=\frac{1000}{\mathrm{Pixels}}\mathrm{\μm}.$

Said step 5):

(1) convert the industrial pure terephthalic acid micro-image into gray level image, selecting the gray-scale value threshold value is 60, is corresponding bianry image with greyscale image transitions, wherein gray-scale value greater than this threshold value be the target area, otherwise be the background area;

(2) capable, the Res row of the 1st row, the every Res of the 1st row beginning from the bianry image matrix scan, and find out the frontier point coordinate and the elemental area S of each industrial pure terephthalic acid particles _Pixel(i), and respectively preserve;

(3) the true area S of each industrial pure terephthalic acid particles _Actual(i) obtain through following formula:

$S_{\mathrm{actual}}\left(i\right)=S_{\mathrm{pixel}}\left(i\right)\×{\left(\frac{1000}{\mathrm{Pixels}}\right)}^2,i=\mathrm{1,2},...,n$

Wherein, S _Actual(i) be the true area of i industrial pure terephthalic acid particles, unit is μ m ²

The present invention proposes that a cover is used for that existing industrial pure terephthalic acid industry size distribution is measured and the low cost of control, the system of high measurement accuracy.Advantages such as it is extensive that this method has application, and sampling amount is few, repeatable high, and data result is abundant

Description of drawings

Fig. 1 finds out all target areas among the figure of source, and obtains the program flow diagram of each target area information simultaneously;

Fig. 2 is through binary conversion treatment, Filtering Processing and marks the micrometer image of each scale mark, numeral 1,6 wherein ..., the 100th, the label of scale mark, " * " of two zone circles represents the coordinate center of scale mark from beginning to end respectively;

Fig. 3 is the image through the industrial pure terephthalic acid of binary conversion treatment, Filtering Processing;

Fig. 4 utilizes program that the industrial pure terephthalic acid image is detected the image that obtains, and wherein the straight line in the industrial pure terephthalic acid particles is the position at each industrial pure terephthalic acid particles maximum particle diameter place;

Fig. 5 is the size distribution figure that utilizes the industrial pure terephthalic acid that several industrial pure terephthalic acid images obtain;

Fig. 6 is the volume distributed median figure that utilizes the industrial pure terephthalic acid that several industrial pure terephthalic acid images obtain;

Fig. 7 is the result who utilizes lognormal distribution that the size distribution match as a result of industrial pure terephthalic acid is obtained;

Fig. 8 is the result who utilizes normal distribution that the volume distributed median match as a result of industrial pure terephthalic acid is obtained.

Embodiment

New particle size distribution estimation method based on the industrial pure terephthalic acid micro-image comprises the steps:

1) stereomicroscope imaging system total magnification is set to 100 times, and wherein the enlargement factor of CCD camera is 25 times, and the enlargement factor of objective lens is 2 times, and regulating zoom and focusing knob enlargement factor is 2 times;

2) micrometer is placed on the objective table, regulate the rising-falling tone focus knob,, utilize computing machine to obtain the true length of each pixel representative in the micrometer image up to observing micrometer image fully clearly;

3) the industrial pure terephthalic acid powder is put into planchet, place vibrating device to handle after 1-2 minute and be transferred on the black backlight of objective table;

4) regulate the rising-falling tone focus knob up to fully clearly observing industrial pure terephthalic acid particles, preserve image to computing machine;

5) be gray level image with the industrial pure terephthalic acid image transitions, carry out binary conversion treatment and Filtering Processing, extract number, true area and the maximum particle diameter of industrial pure terephthalic acid particles in the image;

6) with the true area S of industrial pure terephthalic acid particles _Actual(i), i=1,2 ..., n converts the equivalent diameter and the volume of each industrial pure terephthalic acid particles into, and computing formula is following:

$S_{\mathrm{actual}}\left(i\right)=\frac{\mathrm{\π}}{4}{D_{\mathrm{actual}}}^2\left(i\right)\⇒D_{\mathrm{actual}}\left(i\right)=\sqrt{\frac{4}{\mathrm{\π}}\×S_{\mathrm{actual}}\left(i\right)}$

$V_{\mathrm{actual}}\left(i\right)=\frac{\mathrm{\π}}{6}{D_{\mathrm{actual}}}^3\left(i\right),$ i＝1，2，...，n

D wherein _Actual(i) and V _Actual(i) be respectively the equivalent diameter and the volume of i industrial pure terephthalic acid particles;

7) repeating step 3) to step 6), obtain the equivalent diameter and the volume of several industrial pure terephthalic acid images;

8) utilize the equivalent diameter of industrial pure terephthalic acid to obtain the size distribution and the volume distributed median of industrial pure terephthalic acid;

9) utilize standardized normal distribution, skewness distribution and lognormal distribution that industrial pure terephthalic acid size distribution and volume distributed median result are carried out match, choose true size distribution and the true volume distribution of the minimum distribution of error sum of squares respectively as industrial pure terephthalic acid.

Said step 2):

(1) micrometer is placed on the white backlight plate, regulates the rising-falling tone focus knob, find after the outer circle of scale mark, seek scale mark and scale mark is moved to visual field central authorities; Regulate the rising-falling tone focus knob up to fully clearly observing each scale mark, preserve image to computing machine;

(2) be gray level image with the image transitions that obtains, selecting the reference gray level threshold value is that 50 pairs of gray level images carry out binary conversion treatment, wherein gray-scale value greater than this threshold value be the background area, otherwise be the target area;

(3) utilize the ordfilt2 function of Matlab that image is carried out Filtering Processing, parameter is chosen as the matrix of 6 row, 1 row;

(4) filtered bianry image is carried out line scanning and column scan, find out each target area, wherein the label of scale mark is K1; K2 ..., Kn; The coordinate figure of each Regional Representative's frontier point is kept in the corresponding matrix, and K1 number the centre coordinate with the Kn scale mark obtains through following formula:

$\stackrel{\‾}{X\left(p\right)}=\frac{1}{m}\underset{i=1}{\overset{m}{\mathrm{\Σ}}}X\_\mathrm{coordinate}(p,i),$ $\stackrel{\‾}{Y\left(p\right)}=\frac{1}{m}\underset{i=1}{\overset{m}{\mathrm{\Σ}}}Y\_\mathrm{coordinate}(p,i),$ p＝K1，Kn

X (p) wherein, Y (p) is the coordinate of pixel in the image, and to be each regional coordinate preserved the horizontal ordinate of scale mark to number X_coordinate to m, and Y_coordinate has preserved the horizontal ordinate of scale mark;

(5) pixel count between K1 scale mark and the Kn scale mark coordinate center obtains through following formula:

$\mathrm{Pixels}=\sqrt{{(\stackrel{\‾}{X\left(K1\right)}-\stackrel{\‾}{X\left(\mathrm{Kn}\right)})}^2+{(\stackrel{\‾}{Y\left(K1\right)}-\stackrel{\‾}{Y\left(\mathrm{Kn}\right)})}^2},$

The true length of each pixel representative obtains through following formula:

$1\mathrm{pixel}=\frac{1}{\mathrm{Pixels}}\mathrm{mm}=\frac{1000}{\mathrm{Pixels}}\mathrm{\μm}.$

Said step 5):

(1) convert the industrial pure terephthalic acid micro-image into gray level image, selecting the gray-scale value threshold value is 60, is corresponding bianry image with greyscale image transitions, wherein gray-scale value greater than this threshold value be the target area, otherwise be the background area;

(2) capable, the Res row of the 1st row, the every Res of the 1st row beginning from the bianry image matrix scan, and find out the frontier point coordinate and the elemental area S of each industrial pure terephthalic acid particles _Pixel(i), and respectively preserve;

(3) the true area S of each industrial pure terephthalic acid particles _Actual(i) obtain through following formula:

$S_{\mathrm{actual}}\left(i\right)=S_{\mathrm{pixel}}\left(i\right)\×{\left(\frac{1000}{\mathrm{Pixels}}\right)}^2,$ i＝1，2，...，n

Wherein, S _Actual(i) be the true area of i industrial pure terephthalic acid particles, unit is μ m ²

Existing industrial pure terephthalic acid product to domestic certain chemical plant carries out size distribution and estimates that detailed process is following:

1) stereomicroscope imaging system total magnification being set is 100 times, and wherein the enlargement factor of CCD camera is 25 times, and the enlargement factor of objective lens is 2 times, and the enlargement factor of regulating the zoom and focusing knob is 2 times;

2) micrometer is placed on the white backlight plate of objective table, regulate the rising-falling tone focus knob, find after the outer circle of scale mark, seek scale mark, and scale mark is moved to visual field central authorities.Regulate the rising-falling tone focus knob up to observing micrometer image fully clearly, preserve image to computing machine; With image transitions is gray level image, and selecting the reference gray level threshold value is that 50 pairs of gray level images carry out binary conversion treatment, wherein gray-scale value greater than this threshold value be the background area, otherwise be the target area; Utilize the ordfilt2 function of Matlab that image is carried out Filtering Processing, parameter is chosen as the matrix of 6 row, 1 row; Through the program in the computing machine filtered bianry image is carried out line scanning and column scan, find out each scale mark target area, the process flow diagram of program is please referring to accompanying drawing 1; The label of detected each scale mark is seen shown in the accompanying drawing 2; Wherein, the label of head and the tail scale mark is No. 1 and No. 105, obtains the centre coordinate of scale mark through following formula:

$\stackrel{\‾}{X\left(p\right)}=\frac{1}{m}\underset{i=1}{\overset{m}{\mathrm{\Σ}}}X\_\mathrm{coordinate}(p,i),$ $\stackrel{\‾}{Y\left(p\right)}=\frac{1}{m}\underset{i=1}{\overset{m}{\mathrm{\Σ}}}Y\_\mathrm{coordinate}(p,i)$ ，p＝1,105

X (p) wherein; Y (p) is the coordinate of pixel in the image; M be the coordinate of No. 1 scale mark and No. 105 scale marks to number, X_coordinate and Y_coordinate are respectively horizontal ordinate and the ordinate that the head and the tail scale mark represent frontier point, obtaining from beginning to end, the scale mark centre coordinate is:

X(1)＝320.5，Y(1)＝127.29，X(105)＝310.25，Y(105)＝536.5

Can obtain the pixel count between these two scale marks through following formula:

$\mathrm{Pixels}=\sqrt{{(\stackrel{\‾}{X\left(1\right)}-\stackrel{\‾}{X\left(105\right)})}^2+{(\stackrel{\‾}{Y\left(1\right)}-\stackrel{\‾}{Y\left(105\right)})}^2}=409.34,$

So utilize following formula can obtain the true length of each pixel representative in the image:

$1\mathrm{pixel}=\frac{1}{\mathrm{Pixels}}\mathrm{mm}=\frac{1000}{\mathrm{Pixels}}\mathrm{\μm}=2.44\mathrm{\μm}.$

3) the industrial pure terephthalic acid powder is put into planchet, place vibrating device to handle after 1-2 minute and be transferred on the black backlight of objective table;

4) regulate the rising-falling tone focus knob up to fully clearly observing industrial pure terephthalic acid particles, preserve image to computing machine;

5) utilize the program in the computing machine to be gray level image to the industrial pure terephthalic acid image transitions; Carry out binary conversion treatment and Filtering Processing; Image after the processing is shown in accompanying drawing 3, and wherein the gray threshold in the binary conversion treatment is chosen as 60, and the parameter matrix of Filtering Processing is chosen as 2 row, 1 row; Utilize program to obtain number, true area and the maximum particle diameter of industrial pure terephthalic acid particles in the image, the process flow diagram of program is please referring to accompanying drawing 1; The testing result of image is shown in accompanying drawing 4, and total number of particles is 67, and blue straight line is represented the maximum particle diameter of each particle;

6) utilize the true area S of industrial pure terephthalic acid particles _Actual(i), i=1,2 ..., n obtains the equivalent diameter and the volume of each particle, and computing formula is following:

$S_{\mathrm{actual}}\left(i\right)=\frac{\mathrm{\π}}{4}{D_{\mathrm{actual}}}^2\left(i\right)\⇒D_{\mathrm{actual}}\left(i\right)=\sqrt{\frac{4}{\mathrm{\π}}\×S_{\mathrm{actual}}\left(i\right)}$

$V_{\mathrm{actual}}\left(i\right)=\frac{\mathrm{\π}}{6}{D_{\mathrm{actual}}}^3\left(i\right),$ i＝1，2，...，n

D wherein _Actual(i) and V _Actual(i) be respectively the equivalent diameter and the volume of i particle, the equivalent diameter and volume, each particle diameter grade number of particles and the corresponding volume that obtain each particle are as shown in the table:

7) repeating step 3) to step 6), obtain the equivalent diameter and the volume of the industrial pure terephthalic acid particles of several industrial pure terephthalic acid images;

8) utilize the equivalent diameter of industrial pure terephthalic acid and size distribution and the volume distributed median that volume obtains industrial pure terephthalic acid, distribution results is shown in accompanying drawing 5 and accompanying drawing 6;

9) utilize normal distribution, skewness to distribute and lognormal distribution is carried out match to the size distribution and the volume distributed median result of industrial pure terephthalic acid; Because the particle greater than 280 μ m in the image all is no more than 1 in every width of cloth image, so do not consider the part of particle diameter during match greater than 280 μ m; The error sum of squares of various match distribution gained is as shown in the table:

Lognormal distribution and normal distribution are carried out the error sum of squares minimum that match obtains to the size distribution and the volume distributed median result of industrial pure terephthalic acid; True size distribution and true volume as industrial pure terephthalic acid distributes respectively, and wherein, the parameter of size distribution is μ=0.52; σ=0.92; The parameter of volume distributed median is μ=2.86, σ=9.53, and fitting result is shown in accompanying drawing 7 and accompanying drawing 8.

Claims (1)

1. the new particle size distribution estimation method based on the industrial pure terephthalic acid micro-image is characterized in that comprising the steps:

1) stereomicroscope imaging system total magnification is set to 100 times, and wherein the enlargement factor of CCD camera is 25 times, and the enlargement factor of objective lens is 2 times, and regulating rising-falling tone focus knob enlargement factor is 2 times;

2) micrometer is placed on the objective table, regulate the rising-falling tone focus knob,, utilize computing machine to obtain the true length of each pixel representative in the micrometer image up to observing micrometer image fully clearly;

3) the industrial pure terephthalic acid powder is put into planchet, place vibrating device to handle after 1-2 minute and be transferred on the black backlight of objective table;

4) regulate the rising-falling tone focus knob up to fully clearly observing industrial pure terephthalic acid particles, preserve image to computing machine;

5) be gray level image with the industrial pure terephthalic acid image transitions, carry out binary conversion treatment and Filtering Processing, extract number, true area and the maximum particle diameter of industrial pure terephthalic acid particles in the image;

6) with the true area S of industrial pure terephthalic acid particles _Actual(i), i=1,2 ..., n converts the equivalent diameter and the volume of each industrial pure terephthalic acid particles into, and computing formula is following:

$S_{\mathrm{actual}}\left(i\right)=\frac{\mathrm{\π}}{4}{D_{\mathrm{actual}}}^2\left(i\right)\⇒D_{\mathrm{actual}}\left(i\right)=\sqrt{\frac{4}{\mathrm{\π}}\×S_{\mathrm{actual}}\left(i\right)}$

$V_{\mathrm{actual}}\left(i\right)=\frac{\mathrm{\π}}{6}{D_{\mathrm{actual}}}^3\left(i\right),$ i＝1，2，...，n

D wherein _Actual(i) and V _Actual(i) be respectively the equivalent diameter and the volume of i industrial pure terephthalic acid particles;

7) repeating step 3) to step 6), obtain the equivalent diameter and the volume of several industrial pure terephthalic acid images;

8) utilize the equivalent diameter of industrial pure terephthalic acid to obtain the size distribution and the volume distributed median of industrial pure terephthalic acid;

9) utilize standardized normal distribution, skewness distribution and lognormal distribution that industrial pure terephthalic acid size distribution and volume distributed median result are carried out match, choose true size distribution and the true volume distribution of the minimum distribution of error sum of squares respectively as industrial pure terephthalic acid;

Said step 2):

(1) micrometer is placed on the white backlight plate, regulates the rising-falling tone focus knob, find after the outer circle of scale mark, seek scale mark and scale mark is moved to visual field central authorities; Regulate the rising-falling tone focus knob up to fully clearly observing each scale mark, preserve image to computing machine;

(2) be gray level image with the image transitions that obtains, selecting the reference gray level threshold value is 50, and gray level image is carried out binary conversion treatment, wherein gray-scale value greater than this threshold value be the background area, otherwise be the target area;

(3) utilize the ordfilt2 function of Matlab that image is carried out Filtering Processing, parameter is chosen as the matrix of 6 row, 1 row;

(4) filtered bianry image is carried out line scanning and column scan, find out each target area, wherein the label of scale mark is K1; K2 ..., Kn; The coordinate figure of each Regional Representative's frontier point is kept in the corresponding matrix, and K1 number the centre coordinate with the Kn scale mark obtains through following formula:

$\stackrel{\‾}{X\left(p\right)}=\frac{1}{m}\underset{i=1}{\overset{m}{\mathrm{\Σ}}}X\_\mathrm{coordinate}(p,i),$ $\stackrel{\‾}{Y\left(p\right)}=\frac{1}{m}\underset{i=1}{\overset{m}{\mathrm{\Σ}}}Y\_\mathrm{coordinate}(p,i),$ p＝K1，…Kn

Wherein is the centre coordinate of p scale mark; M is that the coordinate of each regional representative frontier point is to number; X_coordinate is the horizontal ordinate of scale mark, and Y_coordinate is the ordinate of scale mark;

(5) pixel count between K1 scale mark and the Kn scale mark coordinate center obtains through following formula:

$\mathrm{Pixels}=\sqrt{{(\stackrel{\‾}{X\left(K1\right)}-\stackrel{\‾}{X\left(\mathrm{Kn}\right)})}^2+{(\stackrel{\‾}{Y\left(K1\right)}-\stackrel{\‾}{Y\left(\mathrm{Kn}\right)})}^2},$

The true length of each pixel representative obtains through following formula:

$1\mathrm{pixel}=\frac{1}{\mathrm{Pixels}}\mathrm{mm}=\frac{1000}{\mathrm{Pixels}}\mathrm{\μm};$

Said step 5):

(1) convert the industrial pure terephthalic acid micro-image into gray level image, selecting the gray-scale value threshold value is 50, is corresponding bianry image with greyscale image transitions, wherein gray-scale value greater than this threshold value be the target area, otherwise be the background area;

(2) capable, the Res row of the 1st row, the every Res of the 1st row beginning from the bianry image matrix scan, and find out the frontier point coordinate and the elemental area S of each industrial pure terephthalic acid particles _Pixel(i), and respectively preserve;

(3) the true area S of each industrial pure terephthalic acid particles _Actual(i) obtain through following formula:

I=1,2 ..., n, wherein, S _Actual(i) be the true area of i industrial pure terephthalic acid particles, unit is μ m ²

Images