CN111738907A - Train pantograph detection method based on binocular calibration and image algorithm - Google Patents
Train pantograph detection method based on binocular calibration and image algorithm Download PDFInfo
- Publication number
- CN111738907A CN111738907A CN202010511836.3A CN202010511836A CN111738907A CN 111738907 A CN111738907 A CN 111738907A CN 202010511836 A CN202010511836 A CN 202010511836A CN 111738907 A CN111738907 A CN 111738907A
- Authority
- CN
- China
- Prior art keywords
- pantograph
- image
- horn
- template
- goat
- 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
- 238000001514 detection method Methods 0.000 title claims abstract description 121
- OKTJSMMVPCPJKN-UHFFFAOYSA-N Carbon Chemical compound [C] OKTJSMMVPCPJKN-UHFFFAOYSA-N 0.000 claims abstract description 72
- 229910052799 carbon Inorganic materials 0.000 claims abstract description 72
- 238000000034 method Methods 0.000 claims abstract description 68
- 230000009466 transformation Effects 0.000 claims abstract description 64
- 238000005299 abrasion Methods 0.000 claims abstract description 23
- 238000013519 translation Methods 0.000 claims abstract description 16
- 241000594310 Randia aculeata Species 0.000 claims abstract description 14
- 241000877344 Strophanthus divaricatus Species 0.000 claims abstract description 4
- 238000004364 calculation method Methods 0.000 claims abstract description 4
- 241000883966 Astrophytum capricorne Species 0.000 claims description 45
- 239000011159 matrix material Substances 0.000 claims description 33
- 230000007547 defect Effects 0.000 claims description 14
- 230000000877 morphologic effect Effects 0.000 claims description 12
- 230000008569 process Effects 0.000 claims description 9
- 238000000605 extraction Methods 0.000 claims description 7
- 238000012545 processing Methods 0.000 claims description 6
- 230000009467 reduction Effects 0.000 claims description 3
- 238000007689 inspection Methods 0.000 description 12
- 238000003384 imaging method Methods 0.000 description 10
- 238000010586 diagram Methods 0.000 description 3
- 238000005259 measurement Methods 0.000 description 3
- 244000082490 Proboscidea louisianica Species 0.000 description 2
- 235000015926 Proboscidea louisianica ssp. fragrans Nutrition 0.000 description 2
- 235000015925 Proboscidea louisianica subsp. louisianica Nutrition 0.000 description 2
- 206010047571 Visual impairment Diseases 0.000 description 2
- 206010063385 Intellectualisation Diseases 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 230000006996 mental state Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T3/00—Geometric image transformations in the plane of the image
- G06T3/14—Transformations for image registration, e.g. adjusting or mapping for alignment of images
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0004—Industrial image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/80—Analysis of captured images to determine intrinsic or extrinsic camera parameters, i.e. camera calibration
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2200/00—Indexing scheme for image data processing or generation, in general
- G06T2200/32—Indexing scheme for image data processing or generation, in general involving image mosaicing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10004—Still image; Photographic image
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30108—Industrial image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30204—Marker
- G06T2207/30208—Marker matrix
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Quality & Reliability (AREA)
- Current-Collector Devices For Electrically Propelled Vehicles (AREA)
- Image Processing (AREA)
Abstract
The invention discloses a train pantograph detection method based on binocular calibration and an image algorithm, which comprises the following steps: acquiring a left half image and a right half image of a pantograph; after the perspective transformation is carried out on the left half-side image and the right half-side image by calling an image splicing calibration method, the translation splicing transformation is carried out to obtain a pantograph image; calling a pantograph shape template to match the type of the pantograph image to obtain a pantograph type detection result; calling the obtained goat horn template of the pantograph type to match goat horns of the left half image and the right half image of the pantograph image to obtain goat horn detection results; carrying out carbon sliding plate abrasion and carbon sliding plate notch detection on the pantograph to obtain a carbon sliding plate detection result; detecting the position and the pose of the pantograph by a coordinate calculation method to obtain a pantograph pose detection result; and acquiring the pantograph type detection result, the cavel detection result, the carbon sliding plate detection result and the pantograph pose detection result of the steps.
Description
Technical Field
The invention relates to the field of rail transit equipment detection, in particular to a train pantograph detection method based on binocular calibration and an image algorithm.
Background
With the current situations of dense running lines, multiple running vehicles, great influence of vehicle faults and the like of urban rail transit, in order to ensure the reliable quality and the safe running of the train pantograph, the inspection of the train pantograph is finished by manual regular inspection at present. However, the pantograph safety guarantee mode of manual inspection may have the following problems:
(1) potential safety hazards exist: for a long time, for the inspection of the pantograph of the train, the inspection is basically finished by means of the overhead operation of an inspector, and the inspection is carried out by methods such as hand touch, hammer inspection, eye viewing, ear listening, nose smelling and the like, so that the operation environment is hard, the inspection quality is influenced by complex factors such as the technical capability, the physical and mental states and the like of the inspector, the reliability requirement of the quality of the pantograph of the train under heavy work is difficult to guarantee, the inspection missing risk possibly exists, and the hidden danger is buried for the driving safety;
(2) the inspection efficiency is low: the train is run in a large-density manner, the workload of checking the pantograph of the train and the labor intensity of inspectors are greatly increased, all detection items of the pantograph are checked only by the inspectors, the time is long, and the overhauling efficiency is low.
The invention aims to identify the type of the pantograph by carrying out splicing calibration and algorithm processing on the pantograph imaging images of the binocular vision imaging system, obtain the result of whether the pantograph of the train has position and size defects, and output the current state of the pantograph and the values of all detection items to a user, thereby realizing the intellectualization of the online detection of the pantograph, saving the labor and improving the detection precision and efficiency.
Disclosure of Invention
The invention aims to solve the technical problem of how to detect the on-line operation pose and size defects of a train pantograph, and aims to provide a train pantograph detection method based on binocular calibration and an image algorithm to solve the problem.
The invention is realized by the following technical scheme:
a train pantograph detection method based on binocular calibration and an image algorithm comprises the following steps:
s1: acquiring a left half image and a right half image of a pantograph;
s2: after the perspective transformation is carried out on the left half-side image and the right half-side image by calling an image splicing calibration method, the translation splicing transformation is carried out to obtain a pantograph image;
s3: calling a pantograph shape template to match the type of the pantograph image to obtain a pantograph type detection result;
s4: when the pantograph type detection result is that the pantograph type is obtained, calling the obtained goat horn template of the pantograph type to match goat horns of the left half image and the right half image of the pantograph image to obtain a goat horn detection result;
s5: when the horn detection result shows that no serious defect exists in the horn, carrying out carbon sliding plate abrasion and carbon sliding plate notch detection on the pantograph to obtain a carbon sliding plate detection result;
s6: detecting the position and the pose of the pantograph by a coordinate calculation method to obtain a pantograph pose detection result;
s7: and acquiring the pantograph type detection result, the cavel detection result, the carbon sliding plate detection result and the pantograph pose detection result of the steps.
According to the method, after images of the binocular vision imaging system are spliced and calibrated, automatic image processing is performed on each measuring position of a pantograph measuring item, and the numerical value of each detection item is automatically obtained, so that the online running state of the pantograph is judged.
Since binocular vision imaging is based on the parallax principle and acquires two images of the object to be measured from different positions using the imaging apparatus, images of the pantograph, i.e., left and right half-side images, are acquired from the left and right sides of the pantograph in step S1 by the binocular vision imaging method;
however, when image recognition and detection are carried out, a panoramic image of a pantograph is recognized, so that a pantograph image for detection is obtained by image splicing of a collected left half image and a collected right half image;
the invention uses the method of shape template matching when identifying and positioning the type of the pantograph, therefore, the matching template of the shape template needs to be made and stored when detecting the pantograph, the matching template of the shape template mainly related in the invention comprises a left pantograph type template and a right pantograph type template used for identifying the pantograph type and a left ram's horn template and a right ram's horn template used for positioning and calculating;
in the image recognition process, the feature points of the image of the part to be detected in the image need to be found, namely the feature points of the image are matched, and the detection is carried out through the deviation of the feature points obtained through matching and the standard points. When the feature matching detection is carried out, firstly, feature points need to be calculated, and then the feature points are matched, wherein available features comprise sift, surf, fast, angular points and the like;
the invention adopts the angular point of the pantograph as the characteristic point to detect whether the pose of the pantograph generates deviation.
The total length of the area is 2L, the detection value of the abrasion is the difference between the minimum value of the distance between the fitted straight line of the upper end surface curve and the fitted straight line of the lower end surface of the abrasion detection area where the carbon sliding plate component is located and a standard value, and the standard value is set according to the actual situation. In order to meet the requirement of stable and high-precision measurement, the upper end surface curve of the abrasion detection area is obtained in a sub-pixel edge contour extraction mode, the fluctuation transformation of the upper end surface contour can be reflected, higher measurement precision can be guaranteed, and the linear fitting of the lower end surface of the abrasion detection area is realized.
Further, the image stitching calibration method comprises the following steps:
the method comprises the steps that a pantograph calibration plate is hung at a photographing position of a pantograph, and then a left calibration plate image and a right calibration plate image of the pantograph calibration plate are obtained;
acquiring all black squares on the left side of the vertical row of the black circular mark points on the left side calibration plate image, acquiring the angular points of all black squares, solving a minimum enclosing quadrangle of all black squares and acquiring four angular point values of the minimum enclosing quadrangle, solving a left side image perspective transformation matrix of the left side calibration plate image from an image coordinate system to a world coordinate system through the four angular point values and a Zhang calibration method, and performing perspective transformation on the left side calibration plate image through the left side image perspective transformation matrix;
after all black squares on the right side of the vertical row of the black circular mark points on the right side calibration plate image are obtained, the corner points of all the black squares are obtained, the minimum enclosing quadrangle of all the black squares is solved, the four corner point values of the minimum enclosing quadrangle are obtained, the right side image perspective transformation matrix of the right side calibration plate image from the image coordinate system to the world coordinate system is solved through the four corner point values and a Zhang calibration method, and the right side calibration plate image is subjected to perspective transformation through the right side image perspective transformation matrix;
respectively acquiring a left side calibration plate black circular mark point vertical column and a right side calibration plate black circular mark point vertical column of the overlapping part of the left side calibration plate image and the right side calibration plate image after perspective transformation, and performing superposition splicing on the left side calibration plate black circular mark point vertical column and the right side calibration plate black circular mark point vertical column after perspective transformation to obtain a final spliced image and a translation transformation matrix of the spliced image.
Further, the S2 includes the following steps:
s21: calling an image splicing calibration method to obtain a left image perspective transformation matrix, a right image perspective transformation matrix and a translation transformation matrix of the left half-side image and the right half-side image;
s22: and performing perspective transformation and translation splicing transformation on the left half-side image and the right half-side image through the left image perspective transformation matrix, the right image perspective transformation matrix and the translation transformation matrix to obtain a pantograph image.
Further, the pantograph includes:
a front bow left cleat, a rear bow left cleat, a left side connector, a pantograph linkage, a carbon slide assembly, a touch screen carbon slide contact point, a right side connector, a rear bow right cleat, and a front bow right cleat.
The characteristics of the horn part of the pantograph are most easily marked as the type of the pantograph, so the invention combines the models of the horn part of the pantograph into templates with different shapes;
further, the pantograph-shaped template used in S3 includes:
a pantograph type left template including a front pantograph left cleat model matched with the front pantograph left cleat, a rear pantograph left cleat model matched with the rear pantograph left cleat, and a left connector model matched with the left connector;
a pantograph type right template including a rear-pantograph right-side cleat model matched with the rear-pantograph right-side cleat, a front-pantograph right-side cleat model matched with the front-pantograph right-side cleat, and a right-side connector model matched with the right-side connector;
the left cleat template comprises a front arch left cleat model matched with the front arch left cleat and a left connecting piece model matched with the left connecting piece;
and the right cleat template comprises a front arch right cleat model matched with the front arch right cleat and a right connecting piece model matched with the right connecting piece.
The method comprises the steps of firstly detecting the type of the pantograph, wherein the adopted pantograph-shaped template comprises shape templates corresponding to various different types of pantographs, so that the left side templates of all types of pantograph types in the pantograph-shaped template are required to be traversed to be matched with the current pantograph image, the left half side image of the pantograph image is selected in the shape template matching area, and the pantograph type with the highest score, namely the highest left side characteristic similarity, obtained by matching all the left side templates of the pantograph types is defined as the current pantograph type. And only when all the pantograph type left templates are not matched with the characteristics of the pantograph images, calling the pantograph type right templates for matching, otherwise, completing the pantograph type matching.
The left half image and the right half image in the pantograph type matching process are the left half image and the right half image of the pantograph image obtained by performing perspective transformation and stitching in step S2, and are the left region and the right region of one image, that is, the left region and the right region of the pantograph image are subjected to matching operation.
Further, the matching process in S3 includes:
traversing and matching the left half side image of the pantograph image by adopting all pantograph type left side templates in the pantograph-shaped templates to obtain all left side feature similarities;
acquiring a pantograph type corresponding to a pantograph type left template with the highest similarity in all the left side feature similarities as a pantograph type detection result of the pantograph image;
when the pantograph type left template is not matched with the features of the left half-side image of the pantograph image, traversing and matching the right half-side image of the pantograph image by adopting all pantograph type right templates in the pantograph-shaped template to obtain all right-side feature similarities;
acquiring a pantograph type corresponding to the right side template with the highest similarity in all the right side feature similarities as a pantograph type detection result of the pantograph image;
and when all the pantograph type left templates in the pantograph shape templates do not match the characteristics of the left half image of the pantograph image and all the pantograph type right templates in the pantograph shape templates do not match the characteristics of the right half image of the pantograph image, judging that the pantograph head parts of the pantograph have abnormity as a pantograph type detection result.
Further, the process of matching the cavel template in S4 includes:
when the pantograph type detection result is that the pantograph type is obtained, calling the obtained left side goat's horn template and the right side goat's horn template of the pantograph type to respectively perform goat's horn matching on the left half-side image and the right half-side image of the pantograph image;
when the left goat's horn template is not matched with the characteristics of the left image half of the pantograph image or the right goat's horn template is not matched with the characteristics of the right image half of the pantograph image, judging that serious defects exist in the goat's horn before the pantograph as goat's horn detection results;
when the left goat's horn template is matched with the characteristics of the left half image of the pantograph image and the right goat's horn template is matched with the characteristics of the right half image of the pantograph image, performing binarization processing on the left half image and the right half image of the pantograph image matched with the left goat's horn template and the right goat's horn template, and calculating the standard deviation of the area, the shape and the gray scale of the goat's horn by a morphological method;
and when the deviation of the standard deviation of the area, the shape and the gray scale of the goat horn part from the expected value is large, taking the defect and the defect part as the goat horn detection result.
Further, the features of the left half image and the right half image include:
an angular point A of a goat's horn on the left side of the front bow, an end point B and an end point E of the carbon sliding plate assembly, a schematic point D of a touch net carbon sliding plate contact point, an angular point F of a goat's horn on the right side of the front bow and a midpoint C of the carbon sliding plate assembly;
further, the pantograph pose detection in S6 includes:
calculating the deviation of the line coordinates of the end point B and the end point E to obtain the vertical inclination deviation of the pose of the pantograph;
obtaining the center offset deviation of the pose of the pantograph by calculating the deviation of the column coordinates of the midpoint C and the schematic point D;
and obtaining the angle deviation of the pose of the pantograph by calculating the angle of a straight line BE formed by an end point B and an end point E of the carbon sliding plate assembly.
Further, the carbon skid abrasion detection of S5 includes:
selecting the same length to the left and right according to the center of the carbon slide assembly as a wear detection area;
obtaining an upper end surface curve of the abrasion detection area in a sub-pixel edge contour extraction mode;
obtaining a fitting straight line of the lower end surface of the abrasion detection area by a multistage gradient straight line fitting method;
and calculating the difference between the minimum value of the fitted straight line distance between the upper end surface curve and the lower end surface of the abrasion detection area and the standard value to obtain the abrasion detection value of the carbon sliding plate.
Further, the multistage gradient straight line fitting method comprises the following steps:
step 1: extracting a detection area of the lower end face of the pantograph carbon slide plate assembly from the image;
step 2: dividing the detection area into N equal parts, wherein the arithmetic operation of each equal part interval is the same;
and step 3: on the left and right central lines of each equal interval, obtaining the position of the maximum gradient transformation point by using a morphological gradient method, drawing a rectangular area by taking the point as the center of the rectangle, taking the left and right length of each equal interval as the left and right length of the rectangle, and taking the upper and lower 5 pixels of the center of the rectangle as the upper and lower width of the rectangle;
and 4, step 4: dividing the rectangle drawn in the step 3 into n small intervals from left to right, obtaining the edge straight line of the image in each small interval by adopting an image sub-pixel edge straight line extraction method for the small rectangle in each small interval, and taking the middle point of the edge straight line in each small interval as the edge point of the interval;
and 5: calculating the remaining intervals according to the method from the step 3 to the step 4, and finally obtaining the edge points of all the intervals;
step 6: performing optimal straight line fitting on all edge points by adopting a least square method;
and 7: calculating the distances from all the edge points to the best fit straight line, removing the point with the maximum distance deviation of 10 percent, and taking the rest points as the edge points;
and 8: and (4) repeating the steps 6 to 7 for 3 times, fitting the finally obtained edge points into an optimal edge straight line by using a least square method, and taking the straight line as a fitting straight line and a BE connecting line of the lower end surface of the carbon sliding plate assembly.
The notch on the carbon sliding plate assembly can be downwards shaped from the upper end surface of the carbon sliding plate assembly, so that the part with the notch seen on the pantograph image can be directly extracted, the detection in the area without the notch obviously and even the detection in all areas of the carbon sliding plate assembly can be avoided, the complexity of operation is increased, and the suspected notch area is directly selected to be processed and judged.
Further, the carbon sled gap detection of S5 includes:
extracting a suspected gap area of the carbon sliding plate assembly by a morphological gray level image binarization method;
performing morphological open operation and close operation on the extracted suspected gap area of the carbon sliding plate assembly to remove interference generated by noise points;
carrying out layered slicing from top to bottom on the suspected gap area of the carbon sliding plate assembly after the interference is removed;
and performing regional skeletonization treatment on each layer region, and judging whether gaps exist by comparing whether each skeleton line shows an equal-scale reduction rule from top to bottom.
Compared with the prior art, the invention has the following advantages and beneficial effects:
the train pantograph detection method based on the binocular calibration and the image algorithm reduces the workload of manual inspection of the train pantograph, shortens the detection time, optimizes the manual maintenance process, reduces the manual labor intensity and further improves the reliability and efficiency of inspection of the train pantograph.
Drawings
The accompanying drawings, which are included to provide a further understanding of the embodiments of the invention and are incorporated in and constitute a part of this application, illustrate embodiment(s) of the invention and together with the description serve to explain the principles of the invention. In the drawings:
FIG. 1 is a flow chart of the detection method of the present invention;
FIG. 2 is a schematic diagram of the calibration plate image stitching calibration method of the present invention;
FIG. 3 is a schematic diagram of the stitching of the left half image and the right half image of the pantograph according to the present invention;
FIG. 4 is a schematic diagram of the elements of the pantograph measurement feature of the present invention;
FIG. 5 is an enlarged view of a portion of the lower end surface of the carbon slide assembly of the present invention;
FIG. 6 is a schematic view of the fitting of the optimal straight line to the lower end surface of the carbon sliding plate according to the present invention.
Reference numbers and their corresponding part names:
1-horn of the left side of the front bow; 2-rear bow left cleat; 3-left side connector; 4-pantograph linkage; 5-a carbon sled assembly; 6-contacting the net; 7-contact net carbon slide contact point; 8-right side connector; 9-horn of the right side of the back bow; 10-horn of the right side of the anterior arch.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is further described in detail below with reference to examples and accompanying drawings, and the exemplary embodiments and descriptions thereof are only used for explaining the present invention and are not meant to limit the present invention.
Example 1
As shown in fig. 1, a method for detecting a pantograph of a train based on a binocular calibration and an image algorithm includes:
s1: acquiring a left half image and a right half image of a pantograph;
s2: after the perspective transformation is carried out on the left half-side image and the right half-side image by calling an image splicing calibration method, the translation splicing transformation is carried out to obtain a pantograph image;
s3: calling a pantograph shape template to match the type of the pantograph image to obtain a pantograph type detection result;
s4: when the pantograph type detection result is that the pantograph type is obtained, calling the obtained goat horn template of the pantograph type to match goat horns of the left half image and the right half image of the pantograph image to obtain a goat horn detection result;
s5: when the horn detection result shows that no serious defect exists in the horn, carrying out carbon sliding plate abrasion and carbon sliding plate notch detection on the pantograph to obtain a carbon sliding plate detection result;
s6: detecting the position and the pose of the pantograph by a coordinate calculation method to obtain a pantograph pose detection result;
s7: and acquiring the pantograph type detection result, the cavel detection result, the carbon sliding plate detection result and the pantograph pose detection result of the steps.
According to the method, after images of the binocular vision imaging system are spliced and calibrated, automatic image processing is performed on each measuring position of a pantograph measuring item, and the numerical value of each detection item is automatically obtained, so that the online running state of the pantograph is judged.
Since binocular vision imaging is based on the parallax principle and acquires two images of the object to be measured from different positions using the imaging apparatus, images of the pantograph, i.e., left and right half-side images, are acquired from the left and right sides of the pantograph in step S1 by the binocular vision imaging method;
however, when image recognition and detection are carried out, a panoramic image of a pantograph is recognized, so that a pantograph image for detection is obtained by image splicing of a collected left half image and a collected right half image;
the image adopted by the calibration method is shown in fig. 2, and further, the image stitching calibration method comprises the following steps:
the method comprises the steps that a pantograph calibration plate is hung at a photographing position of a pantograph, and then a left calibration plate image and a right calibration plate image of the pantograph calibration plate are obtained;
acquiring all black squares on the left side of the vertical row of the black circular mark points on the left side calibration plate image, acquiring the angular points of all black squares, solving a minimum enclosing quadrangle of all black squares and acquiring four angular point values of the minimum enclosing quadrangle, solving a left side image perspective transformation matrix of the left side calibration plate image from an image coordinate system to a world coordinate system through the four angular point values and a Zhang calibration method, and performing perspective transformation on the left side calibration plate image through the left side image perspective transformation matrix;
after all black squares on the right side of the vertical row of the black circular mark points on the right side calibration plate image are obtained, the corner points of all the black squares are obtained, the minimum enclosing quadrangle of all the black squares is solved, the four corner point values of the minimum enclosing quadrangle are obtained, the right side image perspective transformation matrix of the right side calibration plate image from the image coordinate system to the world coordinate system is solved through the four corner point values and a Zhang calibration method, and the right side calibration plate image is subjected to perspective transformation through the right side image perspective transformation matrix;
respectively acquiring a left side calibration plate black circular mark point vertical column and a right side calibration plate black circular mark point vertical column of the overlapping part of the left side calibration plate image and the right side calibration plate image after perspective transformation, and performing superposition splicing on the left side calibration plate black circular mark point vertical column and the right side calibration plate black circular mark point vertical column after perspective transformation to obtain a final spliced image and a translation transformation matrix of the spliced image.
Further, the S2 includes the following steps:
s21: calling an image splicing calibration method to obtain a left image perspective transformation matrix, a right image perspective transformation matrix and a translation transformation matrix of the left half-side image and the right half-side image;
s22: and performing perspective transformation and translation splicing transformation on the left half-side image and the right half-side image through the left image perspective transformation matrix, the right image perspective transformation matrix and the translation transformation matrix to obtain a pantograph image.
As shown in fig. 3, further, the pantograph includes:
front bow left side goat's horn 1, back bow left side goat's horn 2, left side connecting piece 3, pantograph connecting rod 4, carbon slide subassembly 5, touch net 6, touch net carbon slide contact point 7, right side connecting piece 8, back bow right side goat's horn 9 and front bow right side goat's horn 10.
The characteristics of the horn part of the pantograph are most easily marked as the type of the pantograph, so the invention combines the models of the horn part of the pantograph into templates with different shapes;
further, the pantograph-shaped template used in S3 includes:
the pantograph type left template comprises a front pantograph left cleat model matched with the front pantograph left cleat 1, a rear pantograph left cleat model matched with the rear pantograph left cleat 2 and a left connecting piece model matched with the left connecting piece 3;
a pantograph type right template including a rear-pantograph right-side cleat model matched with the rear-pantograph right-side cleat 9, a front-pantograph right-side cleat model matched with the front-pantograph right-side cleat 10, and a right-side connector model matched with the right-side connector 8;
the left cleat template comprises a front arch left cleat model matched with the front arch left cleat 1 and a left connecting piece model matched with the left connecting piece 3;
and the right cleat template comprises a front arch right cleat model matched with the front arch right cleat 10 and a right connecting piece model matched with the right connecting piece 8.
Further, the matching process in S3 includes:
traversing and matching the left half side image of the pantograph image by adopting all pantograph type left side templates in the pantograph-shaped templates to obtain all left side feature similarities;
acquiring a pantograph type corresponding to a pantograph type left template with the highest similarity in all the left side feature similarities as a pantograph type detection result of the pantograph image;
when the pantograph type left template is not matched with the features of the left half-side image of the pantograph image, traversing and matching the right half-side image of the pantograph image by adopting all pantograph type right templates in the pantograph-shaped template to obtain all right-side feature similarities;
acquiring a pantograph type corresponding to the right side template with the highest similarity in all the right side feature similarities as a pantograph type detection result of the pantograph image;
and when all the pantograph type left templates in the pantograph shape templates do not match the characteristics of the left half image of the pantograph image and all the pantograph type right templates in the pantograph shape templates do not match the characteristics of the right half image of the pantograph image, judging that the pantograph head parts of the pantograph have abnormity as a pantograph type detection result.
Further, the process of matching the cavel template in S4 includes:
when the pantograph type detection result is that the pantograph type is obtained, calling the obtained left side goat's horn template and the right side goat's horn template of the pantograph type to respectively perform goat's horn matching on the left half-side image and the right half-side image of the pantograph image;
when the left goat's horn template is not matched with the characteristics of the left image half of the pantograph image or the right goat's horn template is not matched with the characteristics of the right image half of the pantograph image, judging that serious defects exist in the goat's horn before the pantograph as goat's horn detection results;
when the left goat's horn template is matched with the characteristics of the left half image of the pantograph image and the right goat's horn template is matched with the characteristics of the right half image of the pantograph image, performing binarization processing on the left half image and the right half image of the pantograph image matched with the left goat's horn template and the right goat's horn template, and calculating the standard deviation of the area, the shape and the gray scale of the goat's horn by a morphological method;
and when the deviation of the standard deviation of the area, the shape and the gray scale of the goat horn part from the expected value is large, taking the defect and the defect part as the goat horn detection result.
As shown in fig. 5, further, the features of the left half image and the right half image include:
an angular point A of the front bow left cleat 1, an end point B and an end point E of the carbon sliding plate assembly 5, a schematic point D of a touch net carbon sliding plate contact point 7, an angular point F of the front bow right cleat 10 and a midpoint C of the carbon sliding plate assembly 5.
The connecting line BE of the end point B and the end point E represents a straight line fitted on the lower end face of the carbon slide plate component 5, the end point B is a projection point from the angular point A to the connecting line BE, namely the connecting line AB of the angular point A and the end point B is perpendicular to the connecting line BE, the end point E is a projection point from the angular point F to the connecting line BE, namely the connecting line FE from the angular point F to the end point E is perpendicular to the connecting line BE, the connecting line CO of the middle point C and the point O is a central line of the connecting line AB and the connecting line FE, and the middle point C is an intersection point of.
Further, the carbon skid abrasion detection of S5 includes:
selecting the same length to the left and right according to the center of the carbon slide assembly as a wear detection area;
obtaining an upper end surface curve of the abrasion detection area in a sub-pixel edge contour extraction mode;
obtaining a fitting straight line of the lower end surface of the abrasion detection area by a multistage gradient straight line fitting method;
and calculating the difference between the minimum value of the fitted straight line distance between the upper end surface curve and the lower end surface of the abrasion detection area and the standard value to obtain the abrasion detection value of the carbon sliding plate.
Further, the multistage gradient straight line fitting method comprises the following steps:
step 1: extracting a detection area of the lower end face of the pantograph carbon slide plate assembly from the image, wherein a partial enlarged view of the detection area is shown in fig. 5;
step 2: as shown in fig. 5, the detection area is divided into 10 equal parts, in this embodiment, 10 equal parts are taken as an example, specifically, the divided parts can be determined according to requirements, a first interval from left to right to a tenth interval represent an interval of each equal part, algorithm operations of the intervals of each equal part are the same, and the first interval is taken as an example below;
and step 3: on the left and right central lines of the first interval, obtaining the position of the maximum gradient transformation point by using a morphological gradient method, drawing a rectangular area by taking the point as the center of the rectangle, taking the left and right length of the first interval as the left and right length of the rectangle, and taking the upper and lower 5 pixels respectively at the center of the rectangle as the upper and lower width of the rectangle;
and 4, step 4: dividing the rectangle drawn in the step 3 into 10 small intervals from left to right, obtaining the edge straight line of the image in each small interval by adopting an image sub-pixel edge straight line extraction method for the small rectangle in each small interval, and then taking the middle point of the edge straight line in each small interval as the edge point in the interval;
and 5: calculating the remaining intervals according to the method from the step 3 to the step 4, and finally obtaining the edge points of all the intervals;
step 6: performing optimal straight line fitting on all edge points by adopting a least square method;
and 7: calculating the distances from all the edge points to the best fit straight line, removing the point with the maximum distance deviation of 10 percent, and taking the rest points as the edge points;
and 8: and (4) repeating the steps 6 to 7 for 3 times, fitting the finally obtained edge points into an optimal edge straight line by using a least square method, and taking the straight line as a fitting straight line and a BE connecting line of the lower end surface of the carbon sliding plate assembly. A schematic of the optimal edge line is shown in fig. 6.
Further, the gap on the carbon sliding plate assembly may be in a V shape downward from the upper end surface of the carbon sliding plate assembly, and thus the carbon sliding plate gap detection of S5 includes:
extracting a suspected gap area of the carbon sliding plate assembly by a morphological gray level image binarization method;
performing morphological open operation and close operation on the extracted suspected gap area of the carbon sliding plate assembly to remove interference generated by noise points;
carrying out layered slicing from top to bottom on the suspected gap area of the carbon sliding plate assembly after the interference is removed;
and performing regional skeletonization treatment on each layer region, and judging whether gaps exist by comparing whether each skeleton line shows an equal-scale reduction rule from top to bottom.
Referring to step S6 in fig. 1, the coordinates of point a and point F in fig. 5 can BE obtained through step S4, the straight line BE in fig. 5 can BE obtained through step S5, and the remaining coordinates of point B, point E and point C can BE calculated by knowing point a, point F and straight line BE.
Therefore, referring to fig. 4, the pantograph pose detection in S6 includes:
calculating the deviation of the line coordinates of the end point B and the end point E to obtain the vertical inclination deviation of the pose of the pantograph;
obtaining the center offset deviation of the pose of the pantograph by calculating the deviation of the column coordinates of the midpoint C and the schematic point D;
the angular deviation of the pantograph attitude is obtained by calculating the angle of a straight line BE formed by the end point B and the end point E of the carbon slide plate assembly 5.
The above-mentioned embodiments are intended to illustrate the objects, technical solutions and advantages of the present invention in further detail, and it should be understood that the above-mentioned embodiments are merely exemplary embodiments of the present invention, and are not intended to limit the scope of the present invention, and any modifications, equivalent substitutions, improvements and the like made within the spirit and principle of the present invention should be included in the scope of the present invention.
Claims (10)
1. A train pantograph detection method based on binocular calibration and an image algorithm is characterized by comprising the following steps:
s1: acquiring a left half image and a right half image of a pantograph;
s2: after the perspective transformation is carried out on the left half-side image and the right half-side image by calling an image splicing calibration method, the translation splicing transformation is carried out to obtain a pantograph image;
s3: calling a pantograph shape template to match the type of the pantograph image to obtain a pantograph type detection result;
s4: when the pantograph type detection result is that the pantograph type is obtained, calling the obtained goat horn template of the pantograph type to match goat horns of the left half image and the right half image of the pantograph image to obtain a goat horn detection result;
s5: when the horn detection result shows that no serious defect exists in the horn, carrying out carbon sliding plate abrasion and carbon sliding plate notch detection on the pantograph to obtain a carbon sliding plate detection result;
s6: detecting the position and the pose of the pantograph by a coordinate calculation method to obtain a pantograph pose detection result;
s7: and acquiring the pantograph type detection result, the cavel detection result, the carbon sliding plate detection result and the pantograph pose detection result of the steps.
2. The train pantograph detection method based on the binocular calibration and the image algorithm according to claim 1, wherein the image splicing calibration method comprises the following steps:
the method comprises the steps that a pantograph calibration plate is hung at a photographing position of a pantograph, and then a left calibration plate image and a right calibration plate image of the pantograph calibration plate are obtained;
acquiring all black squares on the left side of the vertical row of the black circular mark points on the left side calibration plate image, acquiring the angular points of all black squares, solving a minimum enclosing quadrangle of all black squares and acquiring four angular point values of the minimum enclosing quadrangle, solving a left side image perspective transformation matrix of the left side calibration plate image from an image coordinate system to a world coordinate system through the four angular point values and a Zhang calibration method, and performing perspective transformation on the left side calibration plate image through the left side image perspective transformation matrix;
after all black squares on the right side of the vertical row of the black circular mark points on the right side calibration plate image are obtained, the corner points of all the black squares are obtained, the minimum enclosing quadrangle of all the black squares is solved, the four corner point values of the minimum enclosing quadrangle are obtained, the right side image perspective transformation matrix of the right side calibration plate image from the image coordinate system to the world coordinate system is solved through the four corner point values and a Zhang calibration method, and the right side calibration plate image is subjected to perspective transformation through the right side image perspective transformation matrix;
respectively acquiring a left side calibration plate black circular mark point vertical column and a right side calibration plate black circular mark point vertical column of the overlapping part of the left side calibration plate image and the right side calibration plate image after perspective transformation, and performing superposition splicing on the left side calibration plate black circular mark point vertical column and the right side calibration plate black circular mark point vertical column after perspective transformation to obtain a final spliced image and a translation transformation matrix of the spliced image.
3. The method for detecting the pantograph of the train based on the binocular calibration and image algorithm as claimed in claim 2, wherein the step S2 includes the steps of:
s21: calling an image splicing calibration method to obtain a left image perspective transformation matrix, a right image perspective transformation matrix and a translation transformation matrix of the left half-side image and the right half-side image;
s22: and performing perspective transformation and translation splicing transformation on the left half-side image and the right half-side image through the left image perspective transformation matrix, the right image perspective transformation matrix and the translation transformation matrix to obtain a pantograph image.
4. The method for detecting the pantograph of the train based on the binocular calibration and the image algorithm, according to claim 1, wherein the pantograph comprises:
a front bow left cleat (1), a rear bow left cleat (2), a left connecting piece (3), a pantograph connecting rod (4), a carbon sliding plate component (5), a touch net (6), a touch net carbon sliding plate contact point (7), a right connecting piece (8), a rear bow right cleat (9) and a front bow right cleat (10);
the pantograph-shaped template used in S3 includes:
the pantograph type left template comprises a front pantograph left cleat model matched with the front pantograph left cleat (1), a rear pantograph left cleat model matched with the rear pantograph left cleat (2) and a left connecting piece model matched with the left connecting piece (3);
a pantograph type right template comprising a rear pantograph right cleat model matched with the rear pantograph right cleat (9), a front pantograph right cleat model matched with the front pantograph right cleat (10) and a right connecting piece model matched with the right connecting piece (8);
the left cleat template comprises a front arch left cleat model matched with the front arch left cleat (1) and a left connecting piece model matched with the left connecting piece (3);
and the right cleat template comprises a front arch right cleat model matched with the front arch right cleat (10) and a right connecting piece model matched with the right connecting piece (8).
5. The method for detecting the pantograph of the train based on the binocular calibration and image algorithm as claimed in claim 4, wherein the matching process in the step S3 includes:
traversing and matching the left half side image of the pantograph image by adopting all pantograph type left side templates in the pantograph-shaped templates to obtain all left side feature similarities;
acquiring a pantograph type corresponding to a pantograph type left template with the highest similarity in all the left side feature similarities as a pantograph type detection result of the pantograph image;
when the pantograph type left template is not matched with the features of the left half-side image of the pantograph image, traversing and matching the right half-side image of the pantograph image by adopting all pantograph type right templates in the pantograph-shaped template to obtain all right-side feature similarities;
acquiring a pantograph type corresponding to the right side template with the highest similarity in all the right side feature similarities as a pantograph type detection result of the pantograph image;
and when all the pantograph type left templates in the pantograph shape templates do not match the characteristics of the left half image of the pantograph image and all the pantograph type right templates in the pantograph shape templates do not match the characteristics of the right half image of the pantograph image, judging that the pantograph head parts of the pantograph have abnormity as a pantograph type detection result.
6. The method for detecting the pantograph of the train based on the binocular calibration and the image algorithm as claimed in claim 4, wherein the goat' S horn template matching process in the step S4 includes:
when the pantograph type detection result is that the pantograph type is obtained, calling the obtained left side goat's horn template and the right side goat's horn template of the pantograph type to respectively perform goat's horn matching on the left half-side image and the right half-side image of the pantograph image;
when the left goat's horn template is not matched with the characteristics of the left image half of the pantograph image or the right goat's horn template is not matched with the characteristics of the right image half of the pantograph image, judging that serious defects exist in the goat's horn before the pantograph as goat's horn detection results;
when the left goat's horn template is matched with the characteristics of the left half image of the pantograph image and the right goat's horn template is matched with the characteristics of the right half image of the pantograph image, performing binarization processing on the left half image and the right half image of the pantograph image matched with the left goat's horn template and the right goat's horn template, and calculating the standard deviation of the area, the shape and the gray scale of the goat's horn by a morphological method;
and when the deviation of the standard deviation of the area, the shape and the gray scale of the goat horn part from the expected value is large, taking the defect and the defect part as the goat horn detection result.
7. The method for detecting the pantograph of the train based on the binocular calibration and image algorithm as claimed in claim 6, wherein the characteristics of the left half-side image and the right half-side image comprise:
an angular point A of a goat's horn (1) on the left side of the front bow, an end point B and an end point E of the carbon sliding plate component (5), a schematic point D of a touch net carbon sliding plate contact point (7), an angular point F of a goat's horn (10) on the right side of the front bow, and a midpoint C of the carbon sliding plate component (5);
the pantograph pose detection in S6 includes:
calculating the deviation of the line coordinates of the end point B and the end point E to obtain the vertical inclination deviation of the pose of the pantograph;
obtaining the center offset deviation of the pose of the pantograph by calculating the deviation of the column coordinates of the midpoint C and the schematic point D;
and obtaining the angle deviation of the pose of the pantograph by calculating the angle of a straight line BE formed by an end point B and an end point E of the carbon sliding plate assembly (5).
8. The method for detecting the pantograph of the train based on the binocular calibration and the image algorithm as claimed in claim 1, wherein the carbon skateboard abrasion detection of S5 comprises:
selecting the same length to the left and right according to the center of the carbon slide assembly as a wear detection area;
obtaining an upper end surface curve of the abrasion detection area in a sub-pixel edge contour extraction mode;
obtaining a fitting straight line of the lower end surface of the abrasion detection area by a multistage gradient straight line fitting method;
and calculating the difference between the minimum value of the fitted straight line distance between the upper end surface curve and the lower end surface of the abrasion detection area and the standard value to obtain the abrasion detection value of the carbon sliding plate.
9. The method for detecting the pantograph of the train based on the binocular calibration and the image algorithm as claimed in claim 8, wherein the multistage gradient straight line fitting method comprises the following steps:
step 1: extracting a detection area of the lower end face of the pantograph carbon slide plate assembly from the image;
step 2: dividing the detection area into N equal parts, wherein the arithmetic operation of each equal part interval is the same;
and step 3: on the left and right central lines of each equal interval, obtaining the position of the maximum gradient transformation point by using a morphological gradient method, drawing a rectangular area by taking the point as the center of the rectangle, taking the left and right length of each equal interval as the left and right length of the rectangle, and taking the upper and lower 5 pixels of the center of the rectangle as the upper and lower width of the rectangle;
and 4, step 4: dividing the rectangle drawn in the step 3 into n small intervals from left to right, obtaining the edge straight line of the image in each small interval by adopting an image sub-pixel edge straight line extraction method for the small rectangle in each small interval, and taking the middle point of the edge straight line in each small interval as the edge point of the interval;
and 5: calculating the remaining intervals according to the method from the step 3 to the step 4, and finally obtaining the edge points of all the intervals;
step 6: performing optimal straight line fitting on all edge points by adopting a least square method;
and 7: calculating the distances from all the edge points to the best fit straight line, removing the point with the maximum distance deviation of 10 percent, and taking the rest points as the edge points;
and 8: and (4) repeating the steps 6 to 7 for 3 times, fitting the finally obtained edge points into an optimal edge straight line by using a least square method, and taking the straight line as a fitting straight line and a BE connecting line of the lower end surface of the carbon sliding plate assembly.
10. The method for detecting the pantograph of the train based on the binocular calibration and the image algorithm as claimed in claim 1, wherein the carbon skateboard gap detection of S5 comprises:
extracting a suspected gap area of the carbon sliding plate assembly by a morphological gray level image binarization method;
performing morphological open operation and close operation on the extracted suspected gap area of the carbon sliding plate assembly to remove interference generated by noise points;
carrying out layered slicing from top to bottom on the suspected gap area of the carbon sliding plate assembly after the interference is removed;
and performing regional skeletonization treatment on each layer region, and judging whether gaps exist by comparing whether each skeleton line shows an equal-scale reduction rule from top to bottom.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010511836.3A CN111738907B (en) | 2020-06-08 | 2020-06-08 | Train pantograph detection method based on binocular calibration and image algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010511836.3A CN111738907B (en) | 2020-06-08 | 2020-06-08 | Train pantograph detection method based on binocular calibration and image algorithm |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111738907A true CN111738907A (en) | 2020-10-02 |
| CN111738907B CN111738907B (en) | 2022-08-23 |
Family
ID=72648461
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010511836.3A Active CN111738907B (en) | 2020-06-08 | 2020-06-08 | Train pantograph detection method based on binocular calibration and image algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111738907B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112781492A (en) * | 2020-12-25 | 2021-05-11 | 江苏集萃智能光电系统研究所有限公司 | Multi-binocular-based train pantograph slide plate abrasion online detection equipment and method |
| CN115810010A (en) * | 2023-01-17 | 2023-03-17 | 广州运达智能科技有限公司 | Train pantograph online detection method and system |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106709905A (en) * | 2016-12-07 | 2017-05-24 | 成都通甲优博科技有限责任公司 | Vibration-proof hammer fault online detection and identification method based on binocular vision image |
| CN107588733A (en) * | 2017-08-21 | 2018-01-16 | 南京理工大学 | A kind of abrasion of pantograph pan on-line measuring device and method based on image |
| CN110455214A (en) * | 2019-07-30 | 2019-11-15 | 中车青岛四方机车车辆股份有限公司 | A kind of abrasion of pantograph pan condition monitoring system and method |
| US20200013144A1 (en) * | 2018-07-04 | 2020-01-09 | Sysmax Innovations Co., Ltd. | Image stitching method and system based on camera earphone |
-
2020
- 2020-06-08 CN CN202010511836.3A patent/CN111738907B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106709905A (en) * | 2016-12-07 | 2017-05-24 | 成都通甲优博科技有限责任公司 | Vibration-proof hammer fault online detection and identification method based on binocular vision image |
| CN107588733A (en) * | 2017-08-21 | 2018-01-16 | 南京理工大学 | A kind of abrasion of pantograph pan on-line measuring device and method based on image |
| US20200013144A1 (en) * | 2018-07-04 | 2020-01-09 | Sysmax Innovations Co., Ltd. | Image stitching method and system based on camera earphone |
| CN110455214A (en) * | 2019-07-30 | 2019-11-15 | 中车青岛四方机车车辆股份有限公司 | A kind of abrasion of pantograph pan condition monitoring system and method |
Non-Patent Citations (1)
| Title |
|---|
| 张新龙: "基于单目视觉激光扫描成像的受电弓磨损检测系统设计", 《CNKI中国知网》 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112781492A (en) * | 2020-12-25 | 2021-05-11 | 江苏集萃智能光电系统研究所有限公司 | Multi-binocular-based train pantograph slide plate abrasion online detection equipment and method |
| CN115810010A (en) * | 2023-01-17 | 2023-03-17 | 广州运达智能科技有限公司 | Train pantograph online detection method and system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111738907B (en) | 2022-08-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109949578B (en) | Vehicle line pressing violation automatic auditing method based on deep learning | |
| CN115482195B (en) | Train part deformation detection method based on three-dimensional point cloud | |
| EP1783683A1 (en) | Mobile peripheral monitor | |
| CN111738907B (en) | Train pantograph detection method based on binocular calibration and image algorithm | |
| CN108921164B (en) | Contact net locator gradient detection method based on three-dimensional point cloud segmentation | |
| JP6369131B2 (en) | Object recognition apparatus and object recognition method | |
| JP5287177B2 (en) | Trolley wire wear and displacement measuring device by image processing | |
| CN113639685B (en) | Displacement detection method, device, equipment and storage medium | |
| CN114627080B (en) | Vehicle stamping accessory defect detection method based on computer vision | |
| CN105258647A (en) | Visual detection method of automobile lock catch rivet point | |
| KR20180115646A (en) | Bead recognition apparatus using vision camera and method thereof | |
| CN116358449A (en) | Aircraft rivet concave-convex amount measuring method based on binocular surface structured light | |
| CN115861274A (en) | Crack detection method integrating three-dimensional point cloud and two-dimensional image | |
| US7893947B2 (en) | Method for extracting edge in photogrammetry with subpixel accuracy | |
| CN110348307B (en) | Path edge identification method and system for crane metal structure climbing robot | |
| CN113624145B (en) | Visual realization method for measuring assembly height and inclination degree of device in automobile fuse box | |
| JP6602595B2 (en) | Vehicle type identification device and vehicle type identification method | |
| CN118196043A (en) | Whole vehicle component detection method based on machine vision and three-dimensional point cloud fusion | |
| CN112651965B (en) | Three-dimensional detection method and system for tread defect of wheel | |
| CN112330667B (en) | Morphology-based laser stripe center line extraction method | |
| CN113947583A (en) | Weld joint nondestructive testing method based on deep learning two-dimensional time sequence image | |
| Khaliluzzaman et al. | Zebra-crossing detection based on geometric feature and vertical vanishing point | |
| TW200949472A (en) | On-board two-dimension contour detection method and system | |
| CN116681912A (en) | Rail gauge detection method and device for railway turnout | |
| CN115931397A (en) | Method and device for determining stable state of rail transit vehicle |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |