CN105021630B - Conveyor belt surface breakage automatic testing method - Google Patents
Conveyor belt surface breakage automatic testing method Download PDFInfo
- Publication number
- CN105021630B CN105021630B CN201510469098.XA CN201510469098A CN105021630B CN 105021630 B CN105021630 B CN 105021630B CN 201510469098 A CN201510469098 A CN 201510469098A CN 105021630 B CN105021630 B CN 105021630B
- Authority
- CN
- China
- Prior art keywords
- max
- conveyor belt
- belt surface
- row
- testing method
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Abstract
The invention discloses a kind of conveyor belt surface breakage automatic testing method, belong to equipment condition monitoring field.The present invention gathers conveyer belt operation image with line array video camera, the uniformity of light intensity is experienced using the same column of image, parted pattern of the row to uneven noise can be suppressed by establishing, realize Fast Segmentation, then edge preserving and noise reduction is carried out to row vector, column vector Mean curve, damaged characteristic parameter is extracted, surface fracture is quickly identified.The conveyor belt surface breakage automatic testing method of the present invention, it is suitable for the damaged on-line checking of conveyor belt surface, conveyor belt surface breakage can be monitored in real time, be easy to repair in time, in case damaged extension, occurs destructive tear.
Description
Technical field
The present invention relates to a kind of conveyor belt surface breakage automatic testing method, more particularly to a kind of based on machine vision
Conveyor belt surface breakage automatic testing method, belongs to equipment condition monitoring field.
Background technology
Conveyer belt is widely used in the industrial circles such as mine, metallurgy, thermal power generation, building materials, chemical industry, industrial harbour
Continuous transport equipment.Conveyer belt by hard foreign matter in the process of running due to being scratched, causing colloid with frame friction, surface area water
The factors such as aging that are hardened influence, conveyor belt surface often occur edge delamination, fester, defect and cracking, hole, rubber cover rouse
The breakage such as bag or peeling, large area abrasion, depth scuffing.Serious surface breakage is to produce the omen of tear, if carrying out not in time
Processing, breakage can grow, extend, and destructive tear finally occurs.At present, the surface fracture maintenance of conveyer belt, is empty in low speed
Carry and human eye detection loading end leaned on by special technical staff under running status, can not detect non-bearing face, frequent start-stop during maintenance
Machine, mechanical wear increase, repair quality and efficiency hardly result in guarantee.It is currently based on the conveyer belt monitoring technology of machine vision also
Laboratory development is in, many key issues are also to be solved, particularly detect the surface fracture of conveyer belt real-time
It is the key technology for needing to be broken through.
The content of the invention
The invention aims to solve to utilize the damaged technical problem of Machine Vision Recognition conveyor belt surface, there is provided one
The automatic testing method of real-time identification surface fracture of the kind based on machine vision.
The technical proposal for solving the technical problem of the invention is：
1st, according to line array video camera characteristic, image Fast Segmentation model is established.
Line array video camera is usually uneven in the light intensity that conveyer belt width is experienced.It is defeated due to mechanical friction
Send belt surface is general in the longitudinal direction to have a slight banding cut, jitter amplitude is small among conveyer belt during operation, both sides shake
Amplitude is big, causes conveyer belt different in the reflected intensity of width light.The illumination of detecting system suffered by other width is very
Hardly possible reaches uniform, particularly curved upper belt.Therefore conveyer belt image be expert at upward intensity profile be often it is very uneven
's.
The uniformity of light intensity is experienced according to same row in the two field picture of line array video camera one, each row gray scale of image is averaged
Value and average intensity suffered by the row are proportional.Ordinary circumstance, response of the object to be measured to light intensity is bigger, non-object to be measured
It is relatively small, therefore can assert that have response to more than half of light intensity is the part relevant with object to be measured, is otherwise exactly
Non- object to be measured.Gray value is considered the portion relevant with object to be measured more than the row average gray half in each row of image
Point, otherwise it is non-object to be measured.Based on principles above, Image Segmentation Model is established.
To the digital picture f (i, j) of conveyer belt, i ＜ H, j ＜ W, H are picture altitude, and W is picture traverse, takes image each column
Average：
Using each column average value by one half as threshold value, image segmentation is carried out：
fout(i, j)=0.5-0.5sgn [f (i, j) -0.5mC(j)+ε)]
Wherein, ε be it is any be more than zero minimum, in order to avoid average and each gray value of the row are zero, at this moment fout
It is not 0, but 0.5.
2nd, using row to standard deviation, interfering between each row light intensity of image is suppressed.
If image each column intensity profile is uneven, also need to consider interfering for light intensity between each row.Each column intensity profile
Uniformity coefficient can be indicated with the row gray variance.
Image each column variance：
Image each column intensity profile evaluation function：
Z (j)=sgn [δC(j)]
If variance δC(j) it is zero, z (j)=0, it is uniform to represent Ben Lie intensity profile, without considering to disturb shadow between column vector
Ring；If δC(j) it is not zero, z (j)=1, represents the row intensity profile less uniformly, it is necessary to consider interference effect between column vector.
Here introduce row, if this value is bigger, illustrates image line to ash to standard deviation, reflection image line to gray scale fluctuations
Big rise and fall is spent, is disturbed between row stronger.
Image line is to standard deviation：
For a constant
In formula,
Row is introduced into image segmentation threshold expression formula, image segmentation threshold to the half of standard deviation as a part for threshold value
For：
T (j)=0.5mC(j)-0.5z(j)·RMSE
Image after segmentation is：
fout(i, j)=0.5-0.5sgn [f (i, j)-T (j)+ε)]
Wherein, ε be it is any be more than zero minimum, in order to avoid threshold value and each gray value of the row are zero, at this moment fout
It is not 0, but 0.5.
3rd, according to row vector, the maximum of column vector average, tentative diagnosis is carried out to breakage.
The average gray of image after segmentation：
Row vector average：
Column vector average：
The maximum of row vector average：Rmax=max (| mR(i)|)
The maximum of column vector average：Cmax=max (| mC′(j)|)
If max (Rmax, Cmax) 255 α of ＞, then there is breakage, be further processed.Otherwise without breakage, damaged inspection is reported
Survey result.Wherein α is adjusting parameter, α ∈ [0.2,0.5].
4th, edge preserving and noise reduction is carried out to row vector, column vector Mean curve, highlights maximum extreme value both sides of the edge yi, i
=1,2 ... N represent row vector, the general symbol(s) of column vector Mean curve, when i takes 3~N-2, in yiNear when taking three
Window：wI, 1=(yi-2, yi-1, yi), wI, 2=(yi-1, yi, yi+1), wI, 3=(yi, yi+1, yi+2), calculate yiWith it is each when window average mk
Deviation (window is numbered when k is, k=1,2,3)：
Select deviationWindow average value when minimum, as i-th point of output valve yout·i。
It is as follows for value at left and right sides of signal：
If yout·i＜ yout·i-1+Th, and yout·i＞ yout·i-1-ThWhen, yout·i=yout·i-1, wherein Th=β max (|
yi|), β is adjusting parameter, β ∈ [0.3,0.5].
5th, surface fracture identifies.
After edge preserving and noise reduction, the Edge Distance of maximum of points both sides in row vector Mean curve is calculated, is exactly damaged
Length Ld；The Edge Distance of maximum of points both sides in column vector Mean curve is calculated, is exactly transverse penetration Wd。
Extract surface fracture characteristic information, including damage length Ld, damaged catercorner length
Finally, surface fracture identification is carried out, criterion of identification is：Damage length Ld＞ γ1H, γ1∈ [0.05,0.1], break
Damage catercorner length
The beneficial effects of the invention are as follows：
The present invention method can from conveyer belt image automatic detection surface fracture, there is intelligent detecting function, it is alternative
The damaged hand inspection of onsite surface, avoids conveyer belt frequent start-stop, reduces mechanical wear, improves detection quality and efficiency.Should
Method has stronger real-time, noise immunity, robustness, is adapted to defeated under the different application occasions such as mine, power generation, metallurgy
Send belt surface breakage on-line checking.
Brief description of the drawings
Fig. 1 is the conveyor belt surface breakage automatic testing method flow chart of the present invention；
Fig. 2 is the surface fracture detection example figure of the present invention；
Fig. 3 is row vector mean analysis instance graph after the image of the present invention is split；
Fig. 4 is column vector mean analysis instance graph after the image of the present invention is split.
Embodiment
In order to better illustrate objects and advantages of the present invention, it is described in further detail below in conjunction with the accompanying drawings.
Conveyor belt surface breakage automatic testing method, its overall technological scheme is as shown in figure 1, specifically comprise the following steps：
Step 1, to the digital picture f (i, j) of conveyer belt, i ＜ H, j ＜ W, H are picture altitude, and W is picture traverse, are calculated
The average m of column vectorC(j), variance δC(j), intensity profile evaluation function z (j) and row vector standard deviation RMSE；
The average m of described column vectorC(j), variance δC(j), intensity profile evaluation function z (j) and row vector standard deviation
RMSECalculating formula be：
(3) z (j)=sgn [δC(j)]；
Wherein
Step 2, segmentation threshold T (j) is calculated, asks for the image f after segmentationout(i, j)；
Image f after described segmentation threshold T (j), segmentationout(i, j) calculation formula is：
(1) T (j)=0.5mC(j)-0.5z(j)·RMSE；
(2)fout(i, j)=0.5-0.5sgn [f (i, j)-T (j)+ε)], ε is any minimum more than zero；
Instance graph 2 (a) size is 1024 × 1024, and the image after segmentation is Fig. 2 (b).
Step 3, the average gray m, row vector average m of image after splitting are asked forR(i), column vector average mC' (j), OK
Vectorial mean-max Rmax, column vector mean-max Cmax；
The average gray m of image after described segmentation, row vector average mR(i), column vector average mC' (j) and row vector
Mean-max Rmax, column vector mean-max CmaxCalculating formula be：
(4)Rmax=max (| mR(i)|)
(5)Cmax=max (| mC′(j)|)
The row vector of instance graph 2 (b), column vector Mean curve are Fig. 3 (a), Fig. 4 (a), Rmax=210.15, Cmax=
11.45。
Step 4, according to the R obtained by step 3max、Cmax, damaged appearance is tentatively determined whether, if not provided, jumping to
Step 8, step 5 is otherwise performed；
Described preliminary judgment criterion is：
If max (Rmax, Cmax) 255 α of ＞, then there is breakage, perform step 5, otherwise without breakage, jump to step 8, its
In, α is adjusting parameter α ∈ [0.2,0.5].
α=0.5 in example, max (Rmax, Cmax) 255 α of ＞, meet damaged condition, perform step 5.
Step 5, using threshold value TR、TCEdge preserving and noise reduction is carried out to row vector, column vector Mean curve respectively, makes step 3
The edge of resulting maximum of points position both sides highlights；
Described threshold value TR、TCCalculating formula be：
(1)TR=β Rmax；
(2)TC=β Cmax；
Wherein, β is adjusting parameter, β ∈ [0.3,0.5]；
β=0.5 in example, Fig. 3 (a), Fig. 4 (a) edge preserving and noise reduction result are Fig. 3 (b), Fig. 4 (b).
Step 6, surface fracture characteristic information is extracted；
Described surface fracture characteristic information includes：Damage length Ld, it is maximum of points both sides in row vector Mean curve
Distance between edge；Damaged catercorner lengthWherein WdIt is transverse penetration, is in column vector Mean curve
Distance between maximum of points both sides of the edge.
Fig. 2 (a) damaged mark is Fig. 2 (c), Ld=79, Wd=39, L=88.
Step 7, the surface fracture characteristic information extracted according to step 6, failure evaluation is carried out；
The criterion of described surface fracture identification is：Damage length Ld＞ γ1H, γ1∈ [0.05,0.1], it is damaged diagonal
Line length
In example, if taking γ1=0.07, γ2=0.05, then meet damaged condition.
Step 8, damage testing result is reported.
Above-described specific descriptions, be to the progress of the purpose, technical scheme and beneficial effect of invention it is further in detail
Explanation.Within the spirit and principles of the invention, any modification, equivalent substitution and improvements done etc., should be included in this hair
Within bright protection domain.
Claims (4)
1. conveyor belt surface breakage automatic testing method, it is characterised in that this method comprises the following steps：
Step 1, to the digital picture f (i, j) of conveyer belt, i ＜ H, j ＜ W, H are picture altitude, and W is picture traverse, calculate row to
The average m of amountC(j), variance δC(j), intensity profile evaluation function z (j) and row vector standard deviation RMSE, calculating formula is respectively
(3) z (j)=sgn [δC(j)],
Wherein
Step 2, segmentation threshold T (j) is calculated, asks for the image f after segmentationout(i, j), calculating formula are respectively
(1) T (j)=0.5mC(j)-0.5z(j)·RMSE,
(2)fout(i, j)=0.5-0.5sgn [f (i, j)-T (j)+ε)], ε is any minimum more than zero；
Step 3, average gray m, the row vector average m of image after splitting are asked forR(i), column vector average mC' (j), row vector
Mean-max Rmax, column vector mean-max Cmax, calculating formula is respectively
(4)Rmax=max (| mR(i) |),
(5)Cmax=max (| mC′(j)|)；
Step 4, according to the R obtained by step 3max、CmaxIf max (Rmax, Cmax) 255 α of ＞, then there is breakage, perform step 5；
Otherwise without breakage, step 8 is jumped to, wherein α is adjusting parameter, α ∈ [0.2,0.5]；
Step 5, using threshold value TR、TCEdge preserving and noise reduction is carried out to row vector, column vector Mean curve respectively, made obtained by step 3
To the edges of maximum of points position both sides highlight；
Step 6, surface fracture characteristic information is extracted；
Step 7, the surface fracture characteristic information extracted according to step 6, failure evaluation is carried out；
Step 8, damage testing result is reported.
2. conveyor belt surface breakage automatic testing method according to claim 1, it is characterised in that in described step 5,
Threshold value TR、TCCalculating formula be
(1)TR=β Rmax,
(2)TC=β Cmax,
Wherein β is adjusting parameter, β ∈ [0.3,0.5].
3. conveyor belt surface breakage automatic testing method according to claim 1, it is characterised in that in described step 6,
Surface fracture characteristic information includes：Damage length Ld, it is the distance in row vector Mean curve between maximum of points both sides of the edge；It is broken
Damage catercorner lengthWherein WdIt is transverse penetration, is maximum of points both sides of the edge in column vector Mean curve
Between distance.
4. conveyor belt surface breakage automatic testing method according to claim 1, it is characterised in that in described step 7,
Surface fracture identification criterion be：Damage length Ld＞ γ1H, γ1∈ [0.05,0.1], damaged catercorner lengthγ2∈ [0.05,0.1].
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510469098.XA CN105021630B (en) | 2015-08-04 | 2015-08-04 | Conveyor belt surface breakage automatic testing method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510469098.XA CN105021630B (en) | 2015-08-04 | 2015-08-04 | Conveyor belt surface breakage automatic testing method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105021630A CN105021630A (en) | 2015-11-04 |
| CN105021630B true CN105021630B (en) | 2017-11-28 |
Family
ID=54411760
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510469098.XA Active CN105021630B (en) | 2015-08-04 | 2015-08-04 | Conveyor belt surface breakage automatic testing method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105021630B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105699391A (en) * | 2016-03-24 | 2016-06-22 | 安徽工程大学 | Detecting device for belt surface of conveyer belt and detection method thereof |
| CN105911074B (en) * | 2016-04-07 | 2018-08-24 | 山西大学 | Adaptive threshold scaling method in wire-core belt lacings X-ray on-line checking |
| CN105823785B (en) * | 2016-05-06 | 2019-06-04 | 西安工业大学 | A kind of conveyer belt alligatoring on-line measuring device and detection method |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US4519041A (en) * | 1982-05-03 | 1985-05-21 | Honeywell Inc. | Real time automated inspection |
| JPH0750039B2 (en) * | 1988-06-08 | 1995-05-31 | 株式会社四国総合研究所 | Conveyor belt damage detector |
| CN102253050A (en) * | 2011-03-14 | 2011-11-23 | 广州市盛通建设工程质量检测有限公司 | Automatic detection method and device for magnetic tile surface defect based on machine vision |
| CN102275723A (en) * | 2011-05-16 | 2011-12-14 | 天津工业大学 | Machine-vision-based online monitoring system and method for conveyer belt |
| CN102565077B (en) * | 2011-11-09 | 2014-07-02 | 天津工业大学 | Method for automatically detecting longitudinal tear of conveyor belt based on machine vision |
-
2015
- 2015-08-04 CN CN201510469098.XA patent/CN105021630B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN105021630A (en) | 2015-11-04 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104914111B (en) | A kind of steel strip surface defect online intelligent recognition detecting system and its detection method | |
| US10783406B1 (en) | Method for detecting structural surface cracks based on image features and bayesian data fusion | |
| CN105021630B (en) | Conveyor belt surface breakage automatic testing method | |
| CN104574389A (en) | Battery piece chromatism selection control method based on color machine vision | |
| CN110490842B (en) | Strip steel surface defect detection method based on deep learning | |
| CN115375686B (en) | Glass edge flaw detection method based on image processing | |
| CN106683099A (en) | Product surface defect detection method | |
| CN106093066A (en) | A kind of magnetic tile surface defect detection method based on the machine vision attention mechanism improved | |
| CN105158258A (en) | Computer vision-based bamboo strip surface defect detection method | |
| CN113658136B (en) | Deep learning-based conveyor belt defect detection method | |
| CN115330780B (en) | Rapid detection method for slag inclusion defect of metal welding | |
| CN114897908B (en) | Machine vision-based method and system for analyzing defects of selective laser powder spreading sintering surface | |
| CN103824304A (en) | Method for performing fault diagnosis on ores on conveying belt during conveying process | |
| CN104637067A (en) | Method for detecting defect of textured surface | |
| CN110827235A (en) | Steel plate surface defect detection method | |
| CN111539927A (en) | Detection process and algorithm of automobile plastic assembly fastening buckle lack-assembly detection device | |
| Sa et al. | Improved Otsu segmentation based on sobel operator | |
| CN114612403A (en) | Intelligent detection method and system for breakage defect of feeding belt | |
| Hashmi et al. | Computer-vision based visual inspection and crack detection of railroad tracks | |
| CN115861294A (en) | Computer vision-based concrete production abnormity detection method and device | |
| CN110349139A (en) | A kind of miscellaneous detection method of packaging based on neural network segmentation | |
| Kleta et al. | Image processing and analysis as a diagnostic tool for an underground infrastructure technical condition monitoring | |
| CN205581025U (en) | Hierarchical detecting system of finish plane bamboo strip vision letter sorting | |
| Aydin et al. | A vision based inspection system using gaussian mixture model based interactive segmentation | |
| Huang et al. | Research on pipe crack detection based on image processing algorithm |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |