CN105588439A - Detection method of sintering machine trolley grate bars - Google Patents
Detection method of sintering machine trolley grate bars Download PDFInfo
- Publication number
- CN105588439A CN105588439A CN201410558164.6A CN201410558164A CN105588439A CN 105588439 A CN105588439 A CN 105588439A CN 201410558164 A CN201410558164 A CN 201410558164A CN 105588439 A CN105588439 A CN 105588439A
- Authority
- CN
- China
- Prior art keywords
- sintering
- image
- grid section
- sintering machine
- sintering pallet
- 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.)
- Granted
Links
Abstract
The invention relates to the field of mineral aggregate sintering for blast furnace process, in particular to a defect detection method of sintering machine trolley grate bars. Provided is a defect detection method of sintering machine trolley grate bars. The method comprises following steps: arranging overlooking images of no-load sintering machine trolleys during collection in no-load regions of sintering machine trolleys during collection of a camera; utilizing overlook images for the sintering machine trolleys to extract bottom surface images of all sintering machine trolleys; setting threshold conditions in order to judge bottom surface images of all sintering machine trolleys; and recognizing defect grate bars.The detection method of sintering machine trolley grate bars has following beneficial effects: by comparing bottom surface images of no-load sintering machine trolleys shoot by the camera and making judgments, grate bars can be automatically monitored to drop down or not; with patrol-inspection, defects and dropping of grate bars can be timely found out in the early stage; without affecting production by dropping grate bars and sintering material, sintering machine trolleys with defected grate bars can be repaired timely; and therefore impact of dropping grate bars and sintering material in great number on sintering production is effectively prevented.
Description
Technical field
The present invention relates to the sintering art of blast furnace process mineral aggregate, relate in particular to a kind of defect detection side of sintering machine bogie gratesMethod.
Background technology
In smelting iron and steel industry, the smelting mineral aggregate that enters blast furnace generally all needs through sintering processes, and sintering machine is to smelt mineral aggregateThe key equipment of sintering process. Sintering machine is in mineral aggregate sintering, and adjacent composition chassis chain of sintering pallet, will prepareGood raw material is loaded on the chassis of sintering machine, then on chassis, loads the surface ignition of raw material, and the bellows of chassis bottom carry outExhausting is combustion-supporting, and along with chassis constantly moves forward, from surface, downward grate firing burns the raw material of being lighted gradually, until whole raw materialGrill thoroughly, lump, complete the sintering of raw material. Sintering machine bogie grates forms after being arranged by the cast iron materials of many rectangular row, dressThe bottom carrying sintered material that is located at chassis carries out sintering, leaves certain gap for ventilative between every grid section, is conducive to materialExhausting in sintering process is combustion-supporting. Because it is wear-resisting, high temperature resistant that sintering machine bogie grates needs, adopt high Cr containing Mn cast iron materialsMoulding by casting, enbrittles, and stressed impact is easily ruptured afterwards. In sintering production process, because of chassis operation, upset,The reasons such as charging, the stressed impact of discharging and high-temperature baking, often can occur to drop after grid section fracture, after indivedual grid section fractures are droppedCan form space, original compact arranged grid section be occurred loosening, after grid section is loosening, there will be inclination, tilt to after certain angleThe hook of grid section both sides can be thrown off support, and therefore a grid section fracture tends to cause whole row's grid section to drop after dropping, and forms oneLarge hole, the sintered material carrying on grid section also drops in bellows below thereupon, and causing sintering machine to leak out affects sintering mineralAmount, the grid section and the raw materials for sintering that drop in a large number in bellows also can threaten to dedusting and the normal operation of exhaust equipment, this situationIf can not find in time to carry out repairing treatment, will normal production of sintering be caused to serious impact.
There is no an applicable detection means due to current, is all to rely on operating personnel to spend a large amount of labours to patrol and examine in industryObserve, even if spent a large amount of labours, be also difficult to find in time that other grid section breaks, and often will wait until that a large amount of grid sections dropWhile forming hole, just can be found, SINTERING PRODUCTION is caused and had a strong impact on. If after finding that bogie grates is damaged, damaged to grid sectionChassis repair, must wait chassis material to unload to fall to running to ad-hoc location and just can carry out, this needs operating personnel to expend greatlyAmount time and efforts is carried out tracking and the location of the damaged chassis of grid section, then could implement to repair, and wastes time and energy.
Summary of the invention
Technical problem to be solved by this invention is to provide sintering machine bogie grates detection method, and the method is by taking video cameraUnloaded sintering pallet bottom picture contrast processing, judge after extracting grid section feature, to the grid section reality that whether dropsShow automatic monitoring; Can find in time in early days the damaged of indivedual grid sections and drop, effectively preventing from causing a large amount of grid sections and raw materials for sinteringThe impact after dropping, SINTERING PRODUCTION being caused.
The present invention is achieved in that a kind of sintering machine bogie grates detection method, takes the photograph at the unloaded region division of sintering palletCamera gathers the sintering pallet overhead view image of operating zero load, then utilizes sintering pallet overhead view image to extract every burningKnot machine trolley ground plan picture, looks like to judge to sintering pallet ground plan after setting threshold condition, identifies damaged grid section.
Build after sintering pallet template and grid section template, comprise the following steps:
S1: the sintering pallet overhead view image that utilizes the unloaded region of camera acquisition;
S2: sintering pallet overhead view image is carried out to pretreatment, intensified-sintered machine trolley feature and grid section feature;
S3: call sintering pallet template and mate with pretreated image, extract sintering pallet ground plan picture;
S4: call grid section template and sintering pallet ground plan looks like to contrast, extract the grid section unit matching; Remaining area isDo not extract region;
S5: setting threshold condition is to each grid section unit and do not extract region and judge, as found to meet the grid section unit of threshold conditionDo not extract region, explanation has grid section disappearance, and now alarm is reported to the police.
In described step S5, described threshold condition, for setting gray threshold, is less than the pixel of gray threshold as lacking taking gray scaleDamage point, the region that damaged point forms is defect area; And set warning length threshold and the warning width threshold value of defect area, whenWhile meeting following two conditions, report to the police simultaneously;
The length of condition one, defect area is more than or equal to warning length threshold;
The width of condition two, defect area is more than or equal to warning width threshold value.
In described step S5, described threshold condition is angle of inclination threshold value, when the direction of grid section unit initial with respect to settingWhen direction deflection exceedes angle of inclination threshold value, report to the police.
Threshold condition in described step S5 comprises threshold value of warning condition and fault threshold condition.
In described structure sintering pallet template and grid section template, the characteristics of image of selecting is edge feature, color characteristic, shapeOne or more any combination in feature, textural characteristics, local invariant feature.
Described step S2 carries out pretreated concrete operations to sintering pallet overhead view image,
The first step, the sintering pallet overhead view image that image recognition arithmetic and control unit is sent image collecting device here is from RGB triple channelColoured image be converted to gray scale single channel image;
Second step, image recognition arithmetic and control unit utilizes filtering algorithm to gray scale single channel image smoothing, denoising;
The 3rd step, the image of image recognition arithmetic and control unit after to smoothing denoising carries out gray-level histogram equalization processing, and usesImage threshold separation algorithms is carried out image binaryzation;
The 4th step, repeatedly iteration is used " corrosion " and " expansion " calculation process, and the image of binaryzation is strengthened, and obtainsBetter sintering pallet and grid section feature.
In described step S3, extract sintering pallet ground plan picture, be specially the edge feature and the part that gather sintering palletConsistency feature, is used Maximun Posterior Probability Estimation Method to mate, and extracts sintering pallet ground plan picture.
In described step S4, extract the grid section unit matching, be specially the edge feature and the shape facility that gather grid section, makeUse based on rotation and mate with convergent-divergent algorithm, extract grid section unit.
Between described step S2 and step S3, also comprise and partially sinter the amalgamation of machine trolley overhead view image and become whole sintering multipleThe step of machine trolley overhead view image.
Sintering machine bogie grates detection method of the present invention contrasts by the unloaded sintering pallet bottom picture that video camera is takenProcess, after extracting grid section feature judge whether grid section is dropped and realized automatic monitoring; Utilize technical side of the present inventionCase, that need not manually require great effort carries out walkaround inspection, just can find in time in early days the damaged of indivedual grid sections and drop, and is not causing combBar and raw materials for sintering drop in a large number to be affected in the situation of producing, and can repair the damaged chassis of grid section in time, effectively preventsThe impact after causing a large amount of grid sections and raw materials for sintering to drop, SINTERING PRODUCTION being caused. Improve sintering work efficiency, alleviate operating personnelThe labour intensity of walkaround inspection, raises labour productivity.
Brief description of the drawings
Fig. 1 is the layout schematic diagram of sintering machine bogie grates detection method device used;
Fig. 2 is the sintering pallet overhead view image in the time that grid section vacancy need to be reported to the police in the present invention;
Fig. 3 is the sintering pallet overhead view image in the time that grid section tilts need to report to the police in the present invention.
In figure: 1 video camera, 2 shooting floor lights, 3 unloaded chassis, 4 feeding devices, 5 fully loaded chassis, 6 detect and prop upFrame, 7 vacancies, 8 inclination grid sections.
Detailed description of the invention
Below in conjunction with specific embodiment, further set forth the present invention. Should be understood that these embodiment are not only for the present invention is describedBe used for limiting the scope of the invention. In addition should be understood that those skilled in the art can after having read the content of the present invention's statementSo that the present invention is made various changes or modifications, these equivalent form of values fall within the model that the application's appended claims limits equallyEnclose.
Embodiment 1
As shown in Figure 1, a kind of sintering machine bogie grates detection method device used comprise bogie grates image camera 1,Shooting floor light 2 and this class main device of image recognition arithmetic and control unit. Taking chassis traffic direction as positive direction, at sintering machineChassis returns to the region of unloaded chassis 3 movable loading point before feeding device 4 chargings become fully loaded chassis 5, on sintering palletAny suitable location of side, can clearly collect bogie grates image, installs one or more Bracket for Inspections 6; PlatformCar grid section image camera, shooting floor light are arranged on checkout gear support, make video camera can photograph chassis in serviceGrid section image. Shooting floor light is followed the collection picture of video camera, makes picture being shot reach certain illumination. Image recognitionArithmetic and control unit can be arranged near the occasion of or other any appropriate, and the output signal of bogie grates image camera is by lookingFrequently cable is connected to the video inputs of image recognition arithmetic and control unit, and the bogie grates image photographing is sent into image recognition fortuneCalculate controller. Adjust bogie grates image camera and make the image frame amplitude of taking comprise a complete bogie grates, ifVisual space is limited can not collect whole bogie grates image in the situation that, also can adopt the multiple bogie grates of multiple supportsThe combination of image camera, finally can splice multiple bogie grates image cameras and covers a complete bogie grates. ?In the case, the grid section image border that make each bogie grates image camera obtain, takes the photograph with adjacent bogie grates imageThe grid section image border that camera obtains, certain overlapping of phase mutual. Adjust the light source of shooting floor light, make it meet shootingMachine is taken the photographing request of bogie grates image, and the identification requirement of image recognition arithmetic and control unit to chassis image.
In the concrete application of the technology of the present invention, bogie grates image camera can adopt arbitrary form to meet grid section IMAQThe video camera requiring. Shooting floor light can adopt the lighting source that arbitrary form is suitable, as LED lamp, fluorescent lamp, halogenLamp etc. Image recognition arithmetic and control unit is a control appliance with microprocessor, its except the input of necessary picture signal andOutside other input for sintering machine control, output signal, being also connected with one can be for the setting of various control parameter and image, numberMan-machine interface according to the show, in the concrete application of the technology of the present invention, image recognition arithmetic and control unit can be that to meet above-mentioned requirements anyThe s operation control equipment of form.
First select characteristics of image to build sintering pallet template and grid section template, in the present embodiment, the characteristics of image of selecting isIn edge feature, color characteristic, shape facility, textural characteristics, local invariant feature (SIFT, SURF, KAZE etc.)One or more any combination, building after sintering pallet template and grid section template sintering machine bogie grates detection method toolBody comprises the following steps:
S1: the sintering pallet overhead view image that utilizes the unloaded region of camera acquisition;
S2: sintering pallet overhead view image is carried out to pretreatment, intensified-sintered machine trolley feature and grid section feature, concrete operations are as follows,
The first step, the sintering pallet overhead view image that image recognition arithmetic and control unit is sent image collecting device here is from RGB triple channelColoured image be converted to gray scale single channel image;
Second step, due to site environment complexity, image can produce much noise in collection and transmission, in order to ensure imageThe height contrast property of edge feature, image recognition arithmetic and control unit utilizes filtering algorithm to carry out image smoothing, denoising.
The 3rd step, in order to adapt to the variation of the brightness of image that the variation of illumination brings, image recognition arithmetic and control unit enters imageThe processing of row gray-level histogram equalization, and use image threshold separation algorithms to carry out image binaryzation.
The 4th step, according to the image after binaryzation is analyzed to discovery, the binary image after denoising still contains much and makes an uproarPoint, and the characteristic of grid section is not obvious, can cause detection error rate higher. Consider above factor, we use mathematicsMorphologic method is processed. In morphology, erosion operation is used and is removed the image detail less than structural element, and expandsComputing is used for interested characteristic in enlarged image, and two kinds of arithmetic operations can ensure not produce overall geometric distortion. IUse erosion operation and dilation operation processing through iteration repeatedly, image is strengthened, obtain better grid section feature.
S2.5: because sintering device handpiece portion back space position is limited, the position of image collecting device is installed with respect to bogie grates positionClose together, video camera vision is limited can not cover all grid sections of whole chassis. For this reason in the present embodiment, adopt multiple figureCarry out distributed image collection as harvester, thereby make video camera vision can cover all grid sections on whole chassis. Based onThe chassis image that distributed image gathers, by region character (spatial domain, frequency domain) and the local characteristic of image adjacent imageMatter (profile, angle point, yardstick invariant features), thus determine the lap position between image, then determine between imageTransformation relation, finally carry out the splicing of image and piece and merge.
In the present embodiment, we carry out the sintering pallet overhead view image of distributed capture by the method based on contour featureSplicing;
S3: call sintering pallet template and mate with gray scale single channel image, extract from the sintering pallet overhead view image of view pictureGo out sintering pallet ground plan picture;
The selection of sintering machine bogie grates image characteristic point is the key point of bogie grates detection system, has determined the standard of chassis locationReally property, also provides basis for follow-up grid section damage condition detects. Before charging, sintering pallet is Light Condition, one of grid sectionBeing engaged togather of root, due to all weather operations of sintering pallet formula, grid section has certain loss, and also can be residual on grid sectionStay material thing; Meanwhile, chassis is ring-type transmission from bottom to up, and in IMAQ, the size of grid section chassis is constantly to change,And there is certain distortion. So the chassis that entirety is made up of grid section is in outward appearance and have in shape very big-difference, this is to platformCar test measuring tape has carried out very large difficulty.
In order to ensure robustness and the repeatability of identification, the application demand detecting according to grid section, it is constant that we select to have yardstickProperty, rotational invariance, illumination invariant image local feature detect operator as foundation characteristic, use series of features vector tableShow the content such as edge, shape of local feature.
Contain significant vertical and horizontal linear edge through the pretreated grid section image of image, as the seam between grid section and grid sectionNeutral gear place between gap, chassis and chassis etc. We can, according to these features, extract chassis feature.
Concrete operation step is as follows:
In the time building sintering pallet template, in order to solve the problem of grid section chassis consistency of scale, we draw in transaction moduleEnter metric space, thereby obtain the grid section chassis image information under different scale by the different running parameter of continuous setup, comprehensiveThese multidate informations, obtain the substantive characteristics of grid section chassis. For grid section chassis image after treatment, extract the dependent office of imagePortion and global characteristics. The feature of extracting can be, the edge feature of image, color characteristic, shape facility (girth, faceLong-pending, prototype degree coefficient, length, width, length-width ratio etc.), textural characteristics (energy, angle second moment, contrast, relevantProperty, uniformity, average gray, gray scale mean square deviation) and local consistency feature (SIFT, SURF, KAZE etc.).
When chassis characteristic model builds, according to the feature distributivity of existing grid section chassis and quantity rank, selection detection modelHypothesis space, then determines method and the strategy learnt; Train and test as relevant by the chassis view data having collectedCard data, select learning strategy according to previous step, select optimized parameter model from hypothesis space, with optimization numerical value meterCalculation method solves relevant parameter, ensures error minimize, and does not occur over-fitting phenomenon.
While using sintering pallet template to mate, the chassis characteristic vector set of obtaining by analysis, by what trainedDetect model of cognition, use Maximun Posterior Probability Estimation Method, determine whether grid section chassis, final, at the bottom of extracting sintering palletFace image.
S4: because working environment more complicated of the present invention can not have lighting condition completely uniformly, so entering in the present inventionBefore the damaged judgement of row, first call grid section template and sintering pallet ground plan looks like to contrast, extract the grid section list matchingUnit; Remaining area is not for extracting region; Respectively to each grid section unit being extracted with do not extract region and judge, to lowerThe probability of false alarm;
According to the feature of grid section and priori, construct the grid section template based on periodic structure in the present invention, this template canThe independent feature based on time domain or frequency domain, combination that also can many features. By the grid section various sizes shape of having takenThe image of shape outward appearance, carries out the model learning of template characteristic, thereby completes the structure of grid section template, utilizes template matches when couplingAlgorithm, these matching algorithms are selected from based on gray scale, based on image pyramid, based on sub-pixel precision, based on rotation and convergent-divergent etc.Algorithm, in searching image, similarity meets the image-region of a certain threshold value; In the present embodiment, extract the grid section list matchingUnit, is specially the edge feature and the shape facility that gather grid section, uses based on rotation and mates with convergent-divergent algorithm, extracts combBar unit.
S5: setting threshold condition is to each grid section unit and do not extract region and judge, find to meet threshold condition grid section unit andDo not extract region, grid section disappearance has been described, alarm is reported to the police; Judge according to two kinds of situations in the present invention,
Grid section on chassis closely links together one by one, and each grid section is shaped as rectangle, and brightness is higher. When grid section has damageTime, the phenomenon presenting on chassis is roughly: grid section excalation, original rectangular-shaped non-rectangle shape that becomes on image; Grid section is completeCome off, on image, still present rectangular-shapedly, but lack reflective therefore this rectangle brightness step-down; After part grid section comes off, periphery is combedBar can produce tilt displacement gradually, fills up grid section defect area. For these features, we design grid section defect detection algorithm.
The first kind as shown in Figure 2, for the vacancy 7 of the damaged rear formation of grid section, is judged vacancy; Described threshold condition is for establishingDetermine gray threshold, the pixel that is less than gray threshold taking gray scale is damaged point, and the region that damaged point forms is defect area; And establishDetermine warning length threshold and the warning width threshold value of defect area, below meeting, when two conditions, report to the police simultaneously;
The length of condition one, defect area is more than or equal to warning length threshold;
The width of condition two, defect area is more than or equal to warning width threshold value.
As shown in Figure 3, although there are some grid sections to drop, there is inclination in all the other grid sections to Equations of The Second Kind, and inclination grid section 8 entirety are filled outMend the area of absence of bulk, therefore utilized this tilt angle alpha to judge; Described threshold condition is angle of inclination threshold value, whenWhen the direction of grid section unit exceedes angle of inclination threshold value with respect to the angle [alpha] of setting inceptive direction deflection, report to the police.
In addition, in the present invention, consider when avoiding false alarm as far as possible, there will not be again and fail to report police, therefore described step S5In threshold condition comprise threshold value of warning condition and fault threshold condition, in the time meeting threshold value of warning condition, will remind operating personnelCheck, in the time meeting fault threshold condition, will send disorderly closedown signal, wait for that operating personnel do further processing;Wherein in the present embodiment, the parameter of threshold value of warning condition be width be 1 grid section width, length be grid section length 2/3, inclineRake angle threshold value is 19 degree; The parameter of fault threshold condition is that width is 1 grid section width, the length length that is grid section, inclinesRake angle threshold value is 30 degree;
When concrete application,
First, intensified-sintered machine trolley feature and grid section feature, be 256 grades of gray spaces by image from color space conversion, logicalCross filtering and process denoising, finally carry out image binaryzation processing, b. processes by the method for mathematical morphology, and 3 times iteration makesProcess with " corrosion " computing and 2 times " expansion ", image is strengthened, obtain better characteristics of image;
Then, image is identified, in the present embodiment, is chosen the space of non-metric space as feature selecting:
The first step, the structure in non-scale feature space. Common characteristic dimension space is linear gaussian pyramid method, thisMethod has been sacrificed the local feature of image, fuzzy minutia. In the present embodiment, the feature space that we choose is non-lineProperty metric space, this model is ensureing under the constant prerequisite in characteristics of image edge, by additive operator splitting algorithm and variable conductionMethod of diffusion, uses compensation arbitrarily, carries out Nonlinear Scale Space Theory tectonic transition.
Second step, grid section chassis feature selecting. In this enforcement, the vertical and horizontal local edge between selection grid section is as spyLevy.
By obtaining grid section related interests characteristic point in the normalized Hessian local maximum of different scale. At integer value chiUnder degree parameter σ, on image, the Hessian defined matrix of p=(x, y) point is: HHessian＝σ2(LxxLyy-L2 xy), whereinLxx (p, σ) expression Gauss second order is ledWith the convolution of image I at p place. Similar have Lxy (p, σ), a Lyy (p, σ).Image recognition arithmetic and control unit, in the process of search local maximum, calculates the detection of each pixel at each layer by above formulaValue, 26 points that are simultaneously adjacent scale domain and image area compare. If the maximum of searching, this value is as grid section platformThe candidate feature point of car.
The 3rd step, grid section chassis characteristic vector represents.
In the present embodiment, we determine the principal direction of characteristic point according to the partial structurtes of characteristic point. When scale parameter is σTime, will detect Dian Wei center, its search radius is 6 σ, all candidate feature points calculating Gausses weighting in it is enclosedSingle order differential value, more connects paracentral contribution margin larger, and the point set using these differential as vector space, taking one as 60 degreeFan-shaped sliding window carry out vector stack, travel through all border circular areas, the longest angle of final result is exactly that this feature is to pointVectorial principal direction.
In the present embodiment, we distribute as the characteristic vector of Expressive Features point using characteristic point neighborhood inside gradient. Image rotating withIts principal direction is alignd, and calculates respectively the little wave response of the 4*4*4 subregion centered by this characteristic pointv＝(Σdx,Σdy,|Σdx|,|Σdy|)tObtain altogether 64 dimension data, last normalization is as the size of the characteristic vector of this characteristic point.
The 4th step, the grid section chassis sample that utilizes multiple different time sections to obtain, maximizes estimation by desired value, calculates and obtainsDesired parameters in grid section chassis Matching Model, thus the Matching Model that realistic site environment uses generated, finally from imageIn extract sintering pallet ground plan picture;
Then, carrying out grid section disappearance detects.
In the present embodiment, we select the rectangular shape of grid section to mate as template characteristic, and based on frequency domain spaceIn, carry out picture search.
A. utilize Fast Fourier Transform (FFT) by image conversion in frequency domain.
B. by conversion after image and its complex conjugate multiply each other, by result again contravariant gain time domain space.
C. in the image after conversion process, the point with higher texture features of brighter value representative image in image. Search officePortion's maximum, this value is as regional center to be detected. Obtain regional center list to be detected.
D. suppose that rectangular configuration is width w and height h, the value for the treatment of surveyed area Center List travels through, interestedRegional extent is w and h, by the similarity of smallest match cost standard of comparison template and template to be detected, finalTake a decision as to whether grid section region.
E. according to the detection sensitivity parameter arranging, for the not searched image-region arriving, if the feature (length and width in regionWith area etc.) while being greater than damaged setting, this region is demarcated as grid section defect area. For the image-region searching, pass throughGray feature threshold value is screened, and in the time exceeding parameter value, this region also will be defined as grid section defect area.
Finally, for all grid section regions that detect, extract the minimum boundary rectangle in grid section region, and calculate rectangle relative level and sitThe anglec of rotation of parameter. According to the decision threshold condition arranging, judge whether this grid section region is grid section defect area.
Claims (10)
1. a sintering machine bogie grates detection method, it is characterized in that: at the sintering pallet overhead view image of the operating zero load of unloaded region division camera acquisition of sintering pallet, then utilize sintering pallet overhead view image to extract every sintering pallet ground plan picture, after setting threshold condition, sintering pallet ground plan is looked like to judge, identify damaged grid section.
2. sintering machine bogie grates detection method as claimed in claim 1, is characterized in that: build after sintering pallet template and grid section template, comprise the following steps:
S1: the sintering pallet overhead view image that utilizes the unloaded region of camera acquisition;
S2: sintering pallet overhead view image is carried out to pretreatment, intensified-sintered machine trolley feature and grid section feature;
S3: call sintering pallet template and mate with image after pretreatment, extract sintering pallet ground plan picture;
S4: call grid section template and sintering pallet ground plan looks like to contrast, extract the grid section unit matching; Remaining area is not for extracting region;
S5: setting threshold condition is to each grid section unit and do not extract region and judge, as found to meet the grid section unit of threshold condition and do not extract region, explanation has grid section to lack, and now alarm is reported to the police.
3. sintering machine bogie grates detection method as claimed in claim 2, is characterized in that: in described step S5, described threshold condition is for setting gray threshold, and the pixel that is less than gray threshold taking gray scale is damaged point, and the region that damaged point forms is defect area; And set warning length threshold and the warning width threshold value of defect area, below meeting, when two conditions, report to the police simultaneously;
The length of condition one, defect area is more than or equal to warning length threshold;
The width of condition two, defect area is more than or equal to warning width threshold value.
4. sintering machine bogie grates detection method as claimed in claim 2, it is characterized in that: in described step S5, described threshold condition is angle of inclination threshold value, in the time that the direction of grid section unit exceedes angle of inclination threshold value with respect to the deflection of setting inceptive direction, reports to the police.
5. the sintering machine bogie grates detection method as described in any claim in claim 2 ~ 4, is characterized in that: the threshold condition in described step S5 comprises threshold value of warning condition and fault threshold condition.
6. the sintering machine bogie grates detection method as described in any claim in claim 2 ~ 4, it is characterized in that: in described structure sintering pallet template and grid section template, the characteristics of image of selecting is one or more any combination in edge feature, color characteristic, shape facility, textural characteristics, local invariant feature.
7. the sintering machine bogie grates detection method as described in any claim in claim 2 ~ 4, is characterized in that: described step S2 carries out pretreated concrete operations to sintering pallet overhead view image and is,
The first step, the sintering pallet overhead view image that image recognition arithmetic and control unit is sent image collecting device here is converted to gray scale single channel image from the three-channel coloured image of RGB;
Second step, image recognition arithmetic and control unit utilizes filtering algorithm to gray scale single channel image smoothing, denoising;
The 3rd step, the image of image recognition arithmetic and control unit after to smoothing denoising carries out gray-level histogram equalization processing, and uses image threshold separation algorithms to carry out image binaryzation;
The 4th step, repeatedly iteration is used " corrosion " and " expansion " calculation process, and the image of binaryzation is strengthened, and obtains better sintering pallet and grid section feature.
8. the sintering machine bogie grates detection method as described in any claim in claim 2 ~ 4, it is characterized in that: in described step S3, extract sintering pallet ground plan picture, be specially the edge feature and the local consistency feature that gather sintering pallet, use Maximun Posterior Probability Estimation Method to mate, extract sintering pallet ground plan picture.
9. the sintering machine bogie grates detection method as described in any claim in claim 2 ~ 4, it is characterized in that: in described step S4, extract the grid section unit matching, be specially the edge feature and the shape facility that gather grid section, use based on rotation and mate with convergent-divergent algorithm, extract grid section unit.
10. the sintering machine bogie grates detection method as described in any claim in claim 2 ~ 4, is characterized in that: between described step S2 and step S3, also comprise and partially sinter the amalgamation of machine trolley overhead view image and become the step of whole sintering pallet overhead view image multiple.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410558164.6A CN105588439B (en) | 2014-10-20 | 2014-10-20 | Sintering machine bogie grates detection method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410558164.6A CN105588439B (en) | 2014-10-20 | 2014-10-20 | Sintering machine bogie grates detection method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105588439A true CN105588439A (en) | 2016-05-18 |
| CN105588439B CN105588439B (en) | 2018-03-09 |
Family
ID=55928206
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410558164.6A Active CN105588439B (en) | 2014-10-20 | 2014-10-20 | Sintering machine bogie grates detection method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105588439B (en) |
Cited By (17)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109489414A (en) * | 2018-11-15 | 2019-03-19 | 北京科技大学 | A kind of continuous atmosphere sintering furnace of high-speed rail brake block |
| CN109724528A (en) * | 2017-10-30 | 2019-05-07 | 中冶长天国际工程有限责任公司 | A kind of determination method and device of the grid section chocking-up degree of sintering pallet |
| CN109724776A (en) * | 2017-10-30 | 2019-05-07 | 中冶长天国际工程有限责任公司 | A kind of determination method and device of the grid section damaged condition of sintering pallet |
| CN110094989A (en) * | 2019-06-14 | 2019-08-06 | 中冶北方（大连）工程技术有限公司 | Roller sintering pallet grate bar falls off detection device |
| CN110094988A (en) * | 2019-06-14 | 2019-08-06 | 中冶北方（大连）工程技术有限公司 | Laser opposite-radiation sintering pallet grate bar falls off detection device and detection method |
| CN110094990A (en) * | 2019-06-14 | 2019-08-06 | 中冶北方（大连）工程技术有限公司 | Reflection type photoelectricity sintering pallet grate bar falls off detection device |
| CN110880003A (en) * | 2019-10-12 | 2020-03-13 | 中国第一汽车股份有限公司 | Image matching method and device, storage medium and automobile |
| CN110991372A (en) * | 2019-12-09 | 2020-04-10 | 河南中烟工业有限责任公司 | Method for identifying cigarette brand display condition of retail merchant |
| CN111223096A (en) * | 2020-03-13 | 2020-06-02 | 中冶长天国际工程有限责任公司 | Method and system for detecting degree of blockage caused by grate bar pasting of sintering machine |
| CN111751898A (en) * | 2020-07-03 | 2020-10-09 | 广东科学技术职业学院 | Device and method for detecting whether core print falls off |
| CN111860240A (en) * | 2020-07-07 | 2020-10-30 | 内蒙古科技大学 | Method and system for detecting offset fault of side plate of trolley of chain grate |
| CN112634220A (en) * | 2020-12-17 | 2021-04-09 | 常州市新创智能科技有限公司 | Breakdown image capturing system for composite material glue filling test |
| CN113570551A (en) * | 2021-07-05 | 2021-10-29 | 天津市三特电子有限公司 | Method and system for monitoring state of grate bars of pellet grate |
| CN113739576A (en) * | 2020-05-28 | 2021-12-03 | 中冶长天国际工程有限责任公司 | Method and system for acquiring section image of tail of sintering machine |
| CN114459245A (en) * | 2022-02-18 | 2022-05-10 | 中冶长天(长沙)智能科技有限公司 | Sintering machine trolley grate bar diagnosis system and method and storage medium |
| EP4050290A1 (en) * | 2021-02-26 | 2022-08-31 | Raytheon Technologies Corporation | A method and system for generating a loaded layout in a vacuum furnace |
| US11620421B2 (en) | 2021-01-29 | 2023-04-04 | General Electric Company | System and method for identifying distortion-compensation threshold for sintering parts with complex features |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH11295040A (en) * | 1998-04-09 | 1999-10-29 | Mitsubishi Electric Corp | Pantograph slider measuring apparatus |
| CN1844899A (en) * | 2005-04-08 | 2006-10-11 | 株式会社名南制作所 | Wide article detection method |
| CN1982904A (en) * | 2005-12-12 | 2007-06-20 | 艾普特佩克股份有限公司 | Apparatus, unit and method for testing image sensor packages |
| CN101027548A (en) * | 2004-06-24 | 2007-08-29 | 埃尔康公司 | Method and apparatus for monitoring and detecting defects in plastic package sealing |
| CN203323558U (en) * | 2013-05-27 | 2013-12-04 | 中冶北方（大连）工程技术有限公司 | Sintering circular cooler trolley |
| CN104101604A (en) * | 2013-04-15 | 2014-10-15 | 宝山钢铁股份有限公司 | Test bench for off-line detection on surface of continuous casting slab |
-
2014
- 2014-10-20 CN CN201410558164.6A patent/CN105588439B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH11295040A (en) * | 1998-04-09 | 1999-10-29 | Mitsubishi Electric Corp | Pantograph slider measuring apparatus |
| CN101027548A (en) * | 2004-06-24 | 2007-08-29 | 埃尔康公司 | Method and apparatus for monitoring and detecting defects in plastic package sealing |
| CN1844899A (en) * | 2005-04-08 | 2006-10-11 | 株式会社名南制作所 | Wide article detection method |
| CN1982904A (en) * | 2005-12-12 | 2007-06-20 | 艾普特佩克股份有限公司 | Apparatus, unit and method for testing image sensor packages |
| CN104101604A (en) * | 2013-04-15 | 2014-10-15 | 宝山钢铁股份有限公司 | Test bench for off-line detection on surface of continuous casting slab |
| CN203323558U (en) * | 2013-05-27 | 2013-12-04 | 中冶北方（大连）工程技术有限公司 | Sintering circular cooler trolley |
Cited By (26)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109724528A (en) * | 2017-10-30 | 2019-05-07 | 中冶长天国际工程有限责任公司 | A kind of determination method and device of the grid section chocking-up degree of sintering pallet |
| CN109724776A (en) * | 2017-10-30 | 2019-05-07 | 中冶长天国际工程有限责任公司 | A kind of determination method and device of the grid section damaged condition of sintering pallet |
| CN109489414A (en) * | 2018-11-15 | 2019-03-19 | 北京科技大学 | A kind of continuous atmosphere sintering furnace of high-speed rail brake block |
| CN110094988A (en) * | 2019-06-14 | 2019-08-06 | 中冶北方（大连）工程技术有限公司 | Laser opposite-radiation sintering pallet grate bar falls off detection device and detection method |
| CN110094989A (en) * | 2019-06-14 | 2019-08-06 | 中冶北方（大连）工程技术有限公司 | Roller sintering pallet grate bar falls off detection device |
| CN110094990A (en) * | 2019-06-14 | 2019-08-06 | 中冶北方（大连）工程技术有限公司 | Reflection type photoelectricity sintering pallet grate bar falls off detection device |
| CN110094988B (en) * | 2019-06-14 | 2024-03-29 | 中冶北方（大连）工程技术有限公司 | Device and method for online detection of grate falling off of trolley of laser correlation sintering machine |
| CN110094989B (en) * | 2019-06-14 | 2024-04-02 | 中冶北方（大连）工程技术有限公司 | Roller type sintering machine trolley grate falling detection device |
| CN110094990B (en) * | 2019-06-14 | 2024-04-02 | 中冶北方（大连）工程技术有限公司 | Reflection-type photoelectric sintering machine trolley grate falling detection device |
| CN110880003A (en) * | 2019-10-12 | 2020-03-13 | 中国第一汽车股份有限公司 | Image matching method and device, storage medium and automobile |
| CN110991372A (en) * | 2019-12-09 | 2020-04-10 | 河南中烟工业有限责任公司 | Method for identifying cigarette brand display condition of retail merchant |
| CN111223096B (en) * | 2020-03-13 | 2024-04-05 | 中冶长天国际工程有限责任公司 | Trolley grate bar paste blocking degree detection method and system of sintering machine |
| CN111223096A (en) * | 2020-03-13 | 2020-06-02 | 中冶长天国际工程有限责任公司 | Method and system for detecting degree of blockage caused by grate bar pasting of sintering machine |
| CN113739576A (en) * | 2020-05-28 | 2021-12-03 | 中冶长天国际工程有限责任公司 | Method and system for acquiring section image of tail of sintering machine |
| CN113739576B (en) * | 2020-05-28 | 2023-06-27 | 中冶长天国际工程有限责任公司 | Method and system for acquiring tail section image of sintering machine |
| CN111751898A (en) * | 2020-07-03 | 2020-10-09 | 广东科学技术职业学院 | Device and method for detecting whether core print falls off |
| CN111860240A (en) * | 2020-07-07 | 2020-10-30 | 内蒙古科技大学 | Method and system for detecting offset fault of side plate of trolley of chain grate |
| CN112634220B (en) * | 2020-12-17 | 2021-08-17 | 常州市新创智能科技有限公司 | Breakdown image capturing system for composite material glue filling test |
| CN112634220A (en) * | 2020-12-17 | 2021-04-09 | 常州市新创智能科技有限公司 | Breakdown image capturing system for composite material glue filling test |
| US11620421B2 (en) | 2021-01-29 | 2023-04-04 | General Electric Company | System and method for identifying distortion-compensation threshold for sintering parts with complex features |
| US11802732B2 (en) | 2021-02-26 | 2023-10-31 | Rtx Corporation | Systems and methods for vacuum furnace post-processing |
| EP4050290A1 (en) * | 2021-02-26 | 2022-08-31 | Raytheon Technologies Corporation | A method and system for generating a loaded layout in a vacuum furnace |
| CN113570551B (en) * | 2021-07-05 | 2023-07-07 | 天津市三特电子有限公司 | Method and system for monitoring grate bar state of pellet grate |
| CN113570551A (en) * | 2021-07-05 | 2021-10-29 | 天津市三特电子有限公司 | Method and system for monitoring state of grate bars of pellet grate |
| CN114459245B (en) * | 2022-02-18 | 2023-06-06 | 中冶长天(长沙)智能科技有限公司 | Sintering machine trolley grate diagnosis system and method and storage medium |
| CN114459245A (en) * | 2022-02-18 | 2022-05-10 | 中冶长天(长沙)智能科技有限公司 | Sintering machine trolley grate bar diagnosis system and method and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105588439B (en) | 2018-03-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105588439A (en) | Detection method of sintering machine trolley grate bars | |
| CN109118482B (en) | Panel defect analysis method and device and storage medium | |
| CN102135416B (en) | Online image detecting system and method for bottle covers | |
| CN104914111B (en) | A kind of steel strip surface defect online intelligent recognition detecting system and its detection method | |
| CN104458755B (en) | Multi-type material surface defect detection method based on machine vision | |
| Mathavan et al. | Use of a self-organizing map for crack detection in highly textured pavement images | |
| Schmugge et al. | Crack segmentation by leveraging multiple frames of varying illumination | |
| CN106991668B (en) | Evaluation method for pictures shot by skynet camera | |
| CN113658136B (en) | Deep learning-based conveyor belt defect detection method | |
| CN110544261B (en) | Method for detecting coal injection state of blast furnace tuyere based on image processing | |
| Quan et al. | The method of the road surface crack detection by the improved Otsu threshold | |
| CN107679495B (en) | Detection method for movable engineering vehicles around power transmission line | |
| CN201935873U (en) | Online image detection system for bottle cap | |
| CN104690007A (en) | Human-machine combined magnetic core on-line detection system and detection method | |
| CN103279765A (en) | Steel wire rope surface damage detection method based on image matching | |
| CN109035226B (en) | Mura defect detection method based on LSTM model | |
| CN108038510A (en) | A kind of detection method based on doubtful flame region feature | |
| CN112395928A (en) | Method for automatically detecting equipment state operation | |
| CN111814678A (en) | Video monitoring-based method and system for identifying coal flow in conveyor belt | |
| CN103509901A (en) | Blast furnace raw fuel particle size detection method | |
| JP6119515B2 (en) | Blast furnace tuyere state observation method and blast furnace tuyere state observation apparatus | |
| CN115908354A (en) | Photovoltaic panel defect detection method based on double-scale strategy and improved YOLOV5 network | |
| Ping et al. | Whole surface defect detection method for bearing rings based on machine vision | |
| Zhang et al. | Fabric defect detection based on visual saliency map and SVM | |
| CN111223098A (en) | Method and system for detecting inclination angle of grate bar of trolley of sintering machine |
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 |