CN107590441A - A kind of pantograph goat's horn on-line measuring device and method based on image procossing - Google Patents
A kind of pantograph goat's horn on-line measuring device and method based on image procossing Download PDFInfo
- Publication number
- CN107590441A CN107590441A CN201710719718.XA CN201710719718A CN107590441A CN 107590441 A CN107590441 A CN 107590441A CN 201710719718 A CN201710719718 A CN 201710719718A CN 107590441 A CN107590441 A CN 107590441A
- Authority
- CN
- China
- Prior art keywords
- mrow
- msubsup
- msub
- horn
- 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.)
- Pending
Links
- 241000883966 Astrophytum capricorne Species 0.000 title claims abstract description 35
- 238000000034 method Methods 0.000 title claims abstract description 25
- 238000001514 detection method Methods 0.000 claims abstract description 65
- 238000012545 processing Methods 0.000 claims abstract description 40
- 230000005540 biological transmission Effects 0.000 claims abstract description 11
- 238000001914 filtration Methods 0.000 claims abstract description 5
- 241000594310 Randia aculeata Species 0.000 claims description 75
- 238000012549 training Methods 0.000 claims description 24
- 230000009466 transformation Effects 0.000 claims description 23
- 241001494479 Pecora Species 0.000 claims description 19
- 239000011159 matrix material Substances 0.000 claims description 15
- 238000005286 illumination Methods 0.000 claims description 12
- 239000013589 supplement Substances 0.000 claims description 9
- 238000013519 translation Methods 0.000 claims description 8
- 238000010276 construction Methods 0.000 claims description 7
- 238000007781 pre-processing Methods 0.000 claims description 7
- 230000001502 supplementing effect Effects 0.000 claims description 7
- 241000877344 Strophanthus divaricatus Species 0.000 claims description 6
- 238000006073 displacement reaction Methods 0.000 claims description 6
- 230000007812 deficiency Effects 0.000 claims description 4
- 238000010845 search algorithm Methods 0.000 claims description 4
- 238000013144 data compression Methods 0.000 claims description 3
- 238000013499 data model Methods 0.000 claims description 3
- 230000002708 enhancing effect Effects 0.000 claims description 3
- 239000000047 product Substances 0.000 claims description 3
- 230000001131 transforming effect Effects 0.000 claims description 3
- 238000012935 Averaging Methods 0.000 claims description 2
- 238000005259 measurement Methods 0.000 abstract description 4
- 230000006641 stabilisation Effects 0.000 abstract 1
- 238000011105 stabilization Methods 0.000 abstract 1
- 238000012360 testing method Methods 0.000 description 13
- 238000010586 diagram Methods 0.000 description 10
- 230000000694 effects Effects 0.000 description 4
- 230000036544 posture Effects 0.000 description 4
- 238000004458 analytical method Methods 0.000 description 3
- 238000010606 normalization Methods 0.000 description 3
- 238000003709 image segmentation Methods 0.000 description 2
- 238000009434 installation Methods 0.000 description 2
- 238000012423 maintenance Methods 0.000 description 2
- 238000005299 abrasion Methods 0.000 description 1
- 238000003556 assay Methods 0.000 description 1
- 230000009194 climbing Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000003745 diagnosis Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
Landscapes
- Image Processing (AREA)
Abstract
The invention discloses a kind of pantograph goat's horn on-line measuring device and method based on image procossing.The device includes image capture module, data transmission module, image processing module three parts, the view data that image capture module gathers is transferred to image processing module by data transmission module, image processing module to real-time capture to pantograph goat's horn imagery exploitation active shape model learning algorithm exist judging goat's horn or missing.Orientation is：First, the image collected is pre-processed, including filtering, image enhaucament, rim detection；Then, characteristic point mark, structure ASM models are carried out to goat's horn image；Then prime area positioning is carried out to goat's horn；Finally positioned, goat's horn is accurately positioned, and judge whether goat's horn lacks according to the goat's horn model of structure and prime area.Convenient, system stabilization that structure of the present invention is laid, can carry out high-precision on line non contact measurement.
Description
Technical Field
The invention relates to the technical field of traffic safety engineering, in particular to an image processing-based pantograph and sheep horn online detection device and method.
Background
The pantograph slide plate is the only contact part of the train and a contact network and is the most important power taking device of the whole train power supply system. In the running process of a train, the pantograph slide plate is continuously contacted with a contact net to cause loss; if the abrasion is serious, the contact net conducting wire clamp can be in a crack or a deep groove on the surface of the sliding plate, so that the pantograph head and the contact net are mechanically collided, and the pantograph fault is caused. Therefore, accurate detection and identification of the pantograph goat's horn in time and effectively are important measures for guaranteeing safety of the pantograph-catenary, so that various safety accidents can be prevented, and a basis is provided for overhauling of the pantograph of the urban rail vehicle.
In the development period of urban rail transit in China, the pantograph detection method is mainly manual detection of manual top climbing. Manual detection needs the train to return to the warehouse and can go on with the contact net outage, and detection efficiency is low, and it is poor to detect the precision, is unfavorable for pantograph maintenance and maintenance. The online detection system of the pantograph is still in a starting stage, and the basic principles adopted at present comprise ultrasonic ranging, CCD imaging, image processing, image identification and the like.
And the thank-you and the like carry out multi-angle shooting on the pantograph by adopting a method of three groups of cameras and one group of cameras after analyzing the actual field installation and detection environment. Although the method can realize the online diagnosis of the fault of the pantograph slide plate, the system has strict requirements on the installation position of key hardware, and the algorithm adopted by the system cannot effectively eliminate the interference of the surrounding environment on the image quality. When the color of the surrounding background environment is close to that of the sliding plate, the upper edge and the lower edge of the pantograph sliding plate cannot be correctly extracted, so that the detection precision is low, and the field practical application is not met.
Disclosure of Invention
The invention aims to provide the image processing-based online detection device and method for the pantograph-sheep horn, which are convenient in structural layout and stable in system, so that high-precision online non-contact measurement is realized.
The technical solution for realizing the purpose of the invention is as follows: the utility model provides a pantograph goat's horn on-line measuring device based on image processing, includes image acquisition module, data transmission module and image processing module, wherein:
the image acquisition module comprises a first wheel axle position sensor, a first photoelectric sensor, a light supplementing device, a camera module, a second photoelectric sensor and a second wheel axle position sensor which are sequentially arranged according to the advancing direction of the train; the camera modules are divided into two groups, each group comprises 2 area-array cameras called half-bow cameras, the half-bow cameras are arranged on the upper side of the roof, a 30-degree overlooking angle is set, and the states of the roof and the pantograph are observed; 2, acquiring pantograph slide plate images from the left direction and the right direction respectively by the planar array camera; the two groups of 4 area array cameras respectively shoot a left half bow in front of the pantograph slide plate, a right half bow in front of the pantograph slide plate, a left half bow in rear of the pantograph slide plate and a right half bow in rear of the pantograph slide plate, and the 4 area array cameras have allowance to shoot a central area of the pantograph slide plate;
the data transmission module is used for transmitting the image data acquired by the image acquisition module to the image processing module;
the image processing module is used for processing the received image data, establishing an active shape model through goat horn sample learning, and judging existence or deficiency of goat horns by utilizing an active shape model learning algorithm for the real-time captured pantograph goat horn images in combination with initial positioning of the pantograph goat horns.
Furthermore, in the image acquisition module, when the first wheel axle position sensor detects a first wheel of the train, the train enters a detection area, and the first photoelectric sensor and the second photoelectric sensor are started simultaneously; when the first photoelectric sensor detects that the pantograph enters a pantograph detection area, starting a light supplementing device to supplement light to an illumination area, enabling the area illumination to meet the photographing requirement, and simultaneously starting a camera module to photograph; when the second photoelectric sensor detects that the pantograph leaves the pantograph detection area, closing the camera module; when the second wheel position sensor detects the 24 th wheel, the train is indicated to leave the detection area, and the image acquisition device and the lighting device in the image acquisition module are turned off.
An image processing-based online detection method for a pantograph-sheep horn comprises the following steps:
step 1, image acquisition: taking a picture by a high-speed camera in an image acquisition module to acquire an original image;
step 2, image preprocessing: filtering, enhancing and detecting the edge of the image;
step 3, the construction method of the goat horn ASM comprises the following steps: the method comprises the steps of calibrating a goat horn learning sample and carrying out ASM training;
step 4, preliminary positioning of a cavel area: carrying out initial positioning on the cavel area;
step 5, horn detection and identification: and matching the goat horn by combining an active shape model learning algorithm, matching the goat horn shape by adopting a single resolution search algorithm, and judging whether the goat horn in the initial positioning area is missing or not.
Further, the cavel ASM construction method in step 3 includes cavel learning sample calibration and ASM training, which are as follows:
(3.1) goat's horn learning sample calibration
After image preprocessing, selecting boundary points and angular points of the goat horn outline as characteristic points, and marking the goat horn characteristic points in a manual mode; in the marking process, the number of characteristic points of each goat horn image is required to be consistent and corresponding to each other, and the shape of the goat horn is described by adopting PDM (product data model), namely the shape of the goat horn image i is represented by all the characteristic points of the goat horn image:
wherein N is the total number of characteristic points of the cavel image;
the cavel image learning sample set is represented as:
wherein M is the total number of the horn images;
(2) ASM training
Firstly, aligning feature points, and specifically comprising the following steps:
a) the shape of a horn xiI is 1,2,3, …, M, and is converted into shape x by translation, rotation and scaling1Align to obtain a transformed shape set
b) Averaging the calculated and transformed cavel images to obtain an average shape m:
wherein,
c) translating, rotating, scaling the average shape m, andaligning;
d) will be provided withCarrying out translation, rotation and scaling transformation, and then aligning and matching with the average shape m;
e) if the average shape is converged, stopping, otherwise, turning to the step b);
the convergence in step e is determined by minimizing the sum of squares of the differences between the aligned horn shapes and the average shape, i.e., finding the transformation TiSo that the following equation is minimized:
∑|m-Ti(xi)|2(4)
the alignment of the horn images is described as: taking two horn shapes as an example, each shape has N coordinate pairs:
firstly, defining transformation matrix T, T is composed of 4 parameters, i.e. rotation angle theta, scale s and translation vector (T)x,ty) Will shape x2And (3) carrying out transformation:
is provided with
Transforming x with T2And x1Alignment, the optimal transformation is obtained by minimizing equation (4):
E＝[x1-Rx2-(tx,ty)T]T(9)
by calculating E pairs of unknown parameters theta, s and tx、tyAnd making a differential equation be zero, so as to solve and obtain a transformation matrix T;
second step, ASM establishment
Obtaining M training shapes after alignment processingEach shape is given by N pairs of coordinates:the average shape is set as:the covariance matrix is then:
wherein S is a 2 Nx 2N matrix;
the variation of the training shape in certain directions is obtained by the eigenvectors of the covariance matrix S, i.e. solving the linear equation:
Spk＝λkpk,k＝1,2,3,…,2N (11)
wherein, the feature vector of S is P, and P is expressed as: p ═ P (P)1,p2,…,p2N)；
For any vector X, there is a shape model parameter b, satisfying:
order:
thus, an estimate of the shape is obtained:
vector btA set of variable model parameters, different b, is definedtDifferent changing shapes can be fitted;
due to biVariance and eigenvalue lambda over the training setiRelated to, biTo satisfy the following equation:
further, the preliminary positioning of the cavel area in the step 4 is as follows:
(1) the collected sheep horns of the pantograph are distributed on the left side and the right side of the image, intersection points are taken as characteristic points, and the left half-bow image and the right half-bow image respectively take the left extension line direction and the right extension line direction of the intersection points as search directions;
(2) the region at least includes two straight lines l1And l2Describing the angle of two straight lines by a straight line methodAndif they are satisfiedThe area is the area to be determined; wherein,andare respectively provided withIs a straight line l1And l2The angle of inclination of (d);
after the area to be determined of the cavel is positioned, because the input image in the cavel detection algorithm is an edge image, curve data compression is directly carried out on edge line segments in the search area on the basis, and then line segments with the length smaller than a set threshold value in the edge line segments are removed; because the directions of the edge line segments of the left and right sheep horns are respectively in the ranges of 35-55 degrees and 125-145 degrees, straight lines are detected by Hough transformation, straight line segments with the angle ranges in the two ranges are searched according to results, and then the inclined edge line segment region is used as an initial positioning region of the sheep horn.
Further, the horn detection and identification in step 5 are as follows:
(1) initializing the goat's horn shape from the average shape in the goat's horn active shape model and the initial position of the goat's horn, as follows:
(2) searching along the boundary normal direction at each mark point of the initial positioning area of the cavel, further obtaining a pixel point with the maximum gradient, marking the point as an optimal target point, moving the mark point to the optimal target point, and if a new target point is not searched, the position of the mark point does not move;
(3) after the mark point is moved, the shape is changed, and a displacement vector exists between the changed shape and the initialized horn shapeAs can be seen from equation (15), the expression after displacement is:
derived from equations (15) and (18):
(4) repeating the steps (2) to (3), and if the preset times of repetition are carried out, p2N,p2N-1…, if the value is less than the threshold value sigma and sigma tends to zero, judging that the goat horn exists, otherwise, judging that the goat horn in the image is missing.
Compared with the prior art, the invention has the remarkable advantages that: (1) the structure is convenient to arrange, the system is stable, and high-precision online non-contact measurement can be performed; (2) can detect the pantograph goat's horn at the train operation in-process, not only guarantee operation safety, improve detection efficiency moreover.
The present invention will be described in further detail below with reference to the accompanying drawings.
Drawings
Fig. 1 is a general system diagram of the on-line detection device for the pantograph horns based on image processing according to the present invention.
Fig. 2 is a flow chart of the on-line detection method of the pantograph sheep horn based on image processing.
FIG. 3 is a graph of the collected horn parts in the present invention, in which (a) to (l) are 12 collected horn parts respectively.
FIG. 4 is a schematic diagram of the characteristic point marking of the goat's horn in the present invention.
Fig. 5 is a schematic diagram of the position of the goat's horn in the image, wherein (a) is a schematic diagram of the left half bow and (b) is a schematic diagram of the right half bow.
FIG. 6 is a graph showing the results of the goat's horn assay of the present invention.
Fig. 7 is a diagram showing the result of the false detection of the cavel in the present invention, in which (a) is a diagram showing the result of the false detection of the cavel with uneven light supplement and (b) is a diagram showing the result of the false detection of the cavel with an excessively large tilt angle.
Detailed Description
The invention relates to an image processing-based pantograph goat horn online detection device and method, which comprises the steps of firstly, filtering and contrast enhancement processing are carried out on an acquired pantograph slider image; and then, preliminarily positioning the goat horn, accurately positioning the goat horn image to be recognized by combining a learned Active Shape Model (ASM) algorithm, judging the existence of the goat horn if the positioning is successful, and otherwise, judging the absence of the goat horn.
The invention is described in further detail below with reference to the figures and the embodiments.
With reference to fig. 1, the present invention provides an image processing-based online detection device for a pantograph-sheep corner, comprising an image acquisition module, a data transmission module and an image processing module, wherein:
the image acquisition module comprises a first wheel axle position sensor, a first photoelectric sensor, a light supplementing device, a camera module, a second photoelectric sensor and a second wheel axle position sensor which are sequentially arranged according to the advancing direction of the train; the camera modules are divided into two groups, each group comprises 2 area-array cameras called half-bow cameras, the half-bow cameras are arranged on the upper side of the roof, a 30-degree overlooking angle is set, and the states of the roof and the pantograph are observed; 2, acquiring pantograph slide plate images from the left direction and the right direction respectively by the planar array camera; the two groups of 4 area array cameras respectively shoot a left half bow in front of the pantograph slide plate, a right half bow in front of the pantograph slide plate, a left half bow in rear of the pantograph slide plate and a right half bow in rear of the pantograph slide plate, and the 4 area array cameras have allowance to shoot a central area of the pantograph slide plate;
the data transmission module is used for transmitting the image data acquired by the image acquisition module to the image processing module;
the image processing module is used for processing the received image data, establishing an active shape model through goat horn sample learning, and judging existence or deficiency of goat horns by utilizing an active shape model learning algorithm for the real-time captured pantograph goat horn images in combination with initial positioning of the pantograph goat horns.
In the image acquisition module, when a first wheel position sensor detects a first wheel of a train, the train enters a detection area, and a first photoelectric sensor and a second photoelectric sensor are started simultaneously; when the first photoelectric sensor detects that the pantograph enters a pantograph detection area, starting a light supplementing device to supplement light to an illumination area, enabling the area illumination to meet the photographing requirement, and simultaneously starting a camera module to photograph; when the second photoelectric sensor detects that the pantograph leaves the pantograph detection area, closing the camera module; when the second wheel position sensor detects the 24 th wheel, the train is indicated to leave the detection area, and the image acquisition device and the lighting device in the image acquisition module are turned off.
With reference to fig. 2, the on-line detection method of the pantograph sheep horn based on image processing of the present invention includes the following steps:
step 1, image acquisition: taking a picture by a high-speed camera in an image acquisition module to acquire an original image;
arranging system devices and acquiring original images: fig. 1 is an overall design diagram of a system. The pantograph goat horn detection system adopts a dynamic non-contact image measurement technology to detect the pantograph goat horn condition. The device mainly comprises an on-site detection device and a control, support and processing device positioned between the devices. The field detection device comprises an image acquisition module, a data transmission module and an image processing module.
The image acquisition module is used for acquiring pantograph images. When the wheel axis sensor 1 detects the first wheel of the train, the train enters a detection area, and meanwhile, the photoelectric sensor is started. When the photoelectric sensor 1 detects that the pantograph enters a pantograph detection area, the light supplementing device is started to supplement light to the illumination area, so that the area illumination meets the photographing requirement, and meanwhile, the high-speed camera is started to photograph. When the photoelectric sensor 2 detects that the pantograph leaves the pantograph detection area, the high-speed camera is turned off. When the 24 th wheel is detected by the wheel axis sensor 2, the train is indicated to leave the detection area, and the image acquisition device and the illumination device are turned off.
The data transmission module is used for transmitting the acquired image data to the data processing module, and the scheme adopts GigE network cable transmission.
And the image processing module is used for processing the image acquired by the image acquisition module. Firstly, carrying out image segmentation on an acquired image, removing a region with useless information and reserving a region with useful information, carrying out image preprocessing such as image denoising and image enhancement on the region with the useful information after the image segmentation, then establishing an active shape model through cavel sample learning, and judging the existence or the deficiency of the cavel of the pantograph cavel image captured in real time by using an active shape model learning algorithm in combination with the initial positioning of the cavel of the pantograph. If the system detects the pantograph fault, the system gives an audible alarm and submits the fault reason so as to maintain the pantograph fault in time and ensure the safety of a pantograph-catenary
Step 2, image preprocessing: filtering, enhancing and detecting the edge of the image;
step 3, the construction method of the goat horn ASM comprises the following steps: the method comprises the steps of calibrating a goat horn learning sample and carrying out ASM training;
the construction method of the goat horn ASM comprises the following steps:
the ASM is based on a Point Distribution Model (PDM), and obtains statistical information of sample feature Point Distribution by training an image sample, and allows a pending target to have a change direction, so as to search for a corresponding feature Point position on a target image.
The construction method of the goat horn ASM comprises the following specific steps:
(3.1) goat's horn learning sample calibration
Fig. 3 is a group of images of the horn portions of 12 collected horn portions (a) to (l). After image preprocessing, boundary points and angular points of the goat horn outline are selected as characteristic points, and the goat horn characteristic points are marked in a manual mode. In the marking process, the number of characteristic points of each goat horn image is required to be consistent and corresponding to each other, and the shape of the goat horn is described by adopting PDM (product data model), namely the shape of the goat horn image i can be represented by all the characteristic points of the goat horn image:
and N is the total number of characteristic points of the goat horn image. The cavel image learning sample set can be expressed as:
wherein M is the total number of the horn images. The schematic diagram of the characteristic point mark of the goat horn is shown in figure 4.
(2) ASM training
The ASM training is divided into two steps, wherein the feature points are aligned in the first step, and the ASM is established in the second step. The feature point alignment is shape normalization, and aims to eliminate non-shape interference of the cavel caused by external factors such as angle, distance, posture transformation and the like in the image, so that the model is more effective. Normalization is generally performed by the general purpose type of analysis of the Procrusts analysis (GPA). The method aligns a series of point distribution models to the same PDM on the basis of not changing the PDM through proper translation, rotation and scaling transformation, so that the acquired original data is not in a disordered state any more, and the interference of non-shape factors is reduced. The alignment process comprises the following steps:
a. the shape of a horn xiI is 1,2,3, …, M, and is translated, rotated, scaled with the shape x1Align to obtain a transformed shape set
b. Calculating the average value of the transformed goat horn images:
wherein,
c. translating, rotating, scaling the average shape m, andaligning;
d. will be provided withCarrying out translation, rotation and scaling transformation, and then aligning and matching with the average shape m;
e. if the average shape converges, stopping, otherwise, turning to step b.
The convergence in step e is determined by minimizing the sum of squares of the differences between the aligned horn shapes and the average shape, i.e., finding the transformation TiSo that the following equation is minimized:
∑|m-Ti(xi)|2(4)
the horn image alignment can be described as: taking two horn shapes as an example, each shape has N coordinate pairs:
firstly, defining transformation matrix T, T is composed of 4 parameters, i.e. rotation angle theta, scale s and translation vector (T)x,ty) Will shape x2And (3) carrying out transformation:
is provided with
Transforming x with T2And x1Alignment, the best transformation can be obtained by minimizing equation (4.37):
E＝[x1-Rx2-(tx,ty)T]T(9)
by calculating E pairs of unknown parameters theta, s and tx、tyAnd making the differential equation zero, thereby solving to obtain a transformation matrix T.
After the horn image shape normalization, an ASM can be established. After the alignment treatment, M training shapes can be obtainedEach shape is given by N pairs of coordinates:the average shape can be set as:the covariance matrix is then:
where S is a 2 Nx 2N matrix.
The variation of the training shape in certain directions can describe important properties of the horn shape, and these properties can be obtained by the eigenvectors of the covariance matrix S, i.e. solving the linear equation:
Spk＝λkpk,k＝1,2,3,…,2N (11)
wherein, the feature vector of S is P, and P can be represented as: p ═ P (P)1,p2,…,p2N)。
For any vector X, there is a shape model parameter b, satisfying:
the vector with large characteristic value is used for describing the direction with large change of the training shape, and when the deviation between the reasonable shape and the average shape is described, p is2N,p2N-1… is negligible, so it is possible to let:
an estimate of the shape can thus be obtained:
if X is a reasonable shape of correlation with the training set, then for a sufficiently large t, the estimate will fit the true shape better.
Vector btA set of variable model parameters, different b, is definedtDifferent varying shapes can be fitted. Due to the fact thatbiVariance and eigenvalue lambda over the training setiIn relation to the preferred shape, biTo satisfy the following equation:
step 4, preliminary positioning of a cavel area: carrying out initial positioning on the cavel area;
the positions of the pantograph sheep horns collected by the 4 CCD cameras are schematically shown in fig. 5(a) - (b) because the pantograph sheep horns are distributed on the left and right sides of the image. The search procedure of the cavel area is as follows:
(1) with the intersection point as a feature point, the left and right half bow images in fig. 5 respectively take the left and right extension line directions of the intersection point as search directions;
(2) the region at least includes two straight lines l1And l2Describing the angle of two straight lines by a straight line methodAndif they are satisfiedThe area is the area to be determined. Wherein,andare respectively a straight line l1And l2The angle of inclination of (a).
After the possible cavel area is located, because the input image in the cavel detection algorithm is the edge image, curve data compression can be directly performed on the edge line segment in the search area on the basis, and then the shorter line segment in the edge line segment is removed. Because the directions of the edge line segments of the left and right sheep horns are respectively in the ranges of 35-55 degrees and 125-145 degrees, straight lines are detected by Hough transformation, straight line segments with the angle ranges in the two ranges are searched according to results, and then the inclined edge line segment region is used as an initial positioning region of the sheep horn.
Step 5, horn detection and identification: and matching the goat horn by combining an active shape model learning algorithm, matching the goat horn shape by adopting a single resolution search algorithm, and judging whether the goat horn in the initial positioning area is missing or not.
The method is characterized in that accurate matching can be carried out on the goat horn by combining an active shape model learning algorithm, the goat horn shape is accurately matched by adopting a single resolution search algorithm, and whether the goat horn really exists in an initial positioning area is further judged, wherein the algorithm comprises the following specific steps:
(1) initializing the goat's horn shape from the average shape in the goat's horn active shape model and the initial position of the goat's horn, as follows:
(2) searching along the boundary normal direction at each mark point initially positioned in a goat horn shape, further acquiring a pixel point with the maximum gradient, marking the point as an optimal target position, moving the mark point to the optimal target point, and if a new target point is not searched, not moving the mark point;
(3) after the mark point is moved, the shape is changed, and a displacement vector exists between the changed shape and the initialized horn shapeAs can be seen from equation (15), the expression after displacement is:
the following equations (15) and (18) can be derived:
(4) repeating the steps (2) and (3), and if the attitude parameter p is repeated for a plurality of times2N,p2N-1And …, if the number of the horn is negligible, judging that the horn exists, otherwise, judging that the horn is absent in the image.
The horn image is identified and positioned by using the algorithm, and the correct detection result display image of the horn is shown in fig. 6.
Example 1
396 cavel images are selected as training sets in the test, the postures of the cavels in the images are different, and the exposure degrees of the images are different. The change of the posture of the goat's horn is about 25 degrees, the image brightness comprises insufficient light supplement, excessive light supplement, uneven light supplement and the like, and the specific proportion of the goat's horn image quality for detection is shown in table 1. Before training a sample, firstly, 396 images are manually calibrated, and each image is calibrated with 28 characteristic points, so that a cavel ASM model is established.
TABLE 1 statistical table of image types
In the test, 264 pictures are selected as a test set to detect the existence of the goat horn, and the images of the training set and the test set are not repeated. The test set images include partial or complete missing goat horn images, uneven illumination images and the like besides the normal goat horn images, and the specific proportion of the test set images is shown in table 2.
TABLE 2 test set image type statistics
The 264 goat horn images are detected by using the goat horn detection algorithm provided by the text, the goat horn recognition time of each picture is about 340ms, 8 real images in the test images have goat horn loss, the test detects that no goat horn exists in 25 images, and the detection result statistical table is shown in table 3. Wherein, 8 images with missing cavel are correctly identified, and the other 17 images with cavel are judged to be missing, and the cavel false-detection rate is 6.4%. Second, the accuracy of the horn identification of the non-ideal image is 79.45%. Wherein, the size of the inclination degree of the goat horn has great influence on the detection accuracy of the goat horn. When the inclination degree of the goat horn is large, the goat horn identification accuracy is only 60%. In addition, the quality of the supplementary lighting effect also has a great influence on the identification accuracy of the cavel, and as can be seen from table 3, the cavel identification accuracy is low under the condition that the supplementary lighting effect of the image is not ideal.
TABLE 3 statistical table of test results
Because this patent adopts and carries out the accurate positioning to the goat's horn based on initiative shape model algorithm, and the illumination influences the testing result less in theory, nevertheless can know from the table that the detection accuracy rate when the light filling effect is unsatisfactory is lower relatively. This is mainly because the gray scale distribution of the goat's horn image is not uniform due to the poor light filling effect. Secondly, the original characteristic of the goat's horn can be strengthened or weakened by illuminating the projected shadow, thereby reducing the accuracy rate of goat's horn recognition. In order to analyze the specific influence of illumination and the cavel inclination angle on cavel identification, two images of false detection in the test are selected, and the image of the false detection result of the cavel is shown in fig. 7. As can be seen from the analysis of fig. 7(a), when the system collects an image, the image in most areas of the horns of the pantograph is dark due to uneven light supplement, and although the image is enhanced, the influence caused by uneven light cannot be effectively compensated. When the active shape model-based cavel positioning method is used for establishing a local gray structure model, the extracted gray value is a derivative value of the gray change of the feature point along the contour normal direction, but not an absolute value, and the influence of external light on the positioning accuracy can be overcome to a certain extent. But because the quality of the test image is poor, the light is uneven, so that partial information is lost, and the detection of the goat's horn is wrong. Fig. 7(b) shows that false detection is caused by an excessively large goat horn inclination angle, although the influence of the goat horn positioning method based on the active shape model on different postures has certain robustness, if the goat horn inclination angle exceeds a certain range, an area matched with the feature point cannot be found in the image, and then the goat horn detection fails.
Claims (6)
1. The utility model provides a pantograph goat's horn on-line measuring device based on image processing which characterized in that, includes image acquisition module, data transmission module and image processing module, wherein:
the image acquisition module comprises a first wheel axle position sensor, a first photoelectric sensor, a light supplementing device, a camera module, a second photoelectric sensor and a second wheel axle position sensor which are sequentially arranged according to the advancing direction of the train; the camera modules are divided into two groups, each group comprises 2 area-array cameras called half-bow cameras, the half-bow cameras are arranged on the upper side of the roof, a 30-degree overlooking angle is set, and the states of the roof and the pantograph are observed; 2, acquiring pantograph slide plate images from the left direction and the right direction respectively by the planar array camera; the two groups of 4 area array cameras respectively shoot a left half bow in front of the pantograph slide plate, a right half bow in front of the pantograph slide plate, a left half bow in rear of the pantograph slide plate and a right half bow in rear of the pantograph slide plate, and the 4 area array cameras have allowance to shoot a central area of the pantograph slide plate;
the data transmission module is used for transmitting the image data acquired by the image acquisition module to the image processing module;
the image processing module is used for processing the received image data, establishing an active shape model through goat horn sample learning, and judging existence or deficiency of goat horns by utilizing an active shape model learning algorithm for the real-time captured pantograph goat horn images in combination with initial positioning of the pantograph goat horns.
2. The device for detecting the angle of a pantograph of a railway vehicle based on image processing as claimed in claim 1, wherein in the image acquisition module, when the first wheel axle position sensor detects the first wheel of the train, it indicates that the train enters the detection area, and the first and second photoelectric sensors are turned on simultaneously; when the first photoelectric sensor detects that the pantograph enters a pantograph detection area, starting a light supplementing device to supplement light to an illumination area, enabling the area illumination to meet the photographing requirement, and simultaneously starting a camera module to photograph; when the second photoelectric sensor detects that the pantograph leaves the pantograph detection area, closing the camera module; when the second wheel position sensor detects the 24 th wheel, the train is indicated to leave the detection area, and the image acquisition device and the lighting device in the image acquisition module are turned off.
3. An image processing-based online detection method for a pantograph-sheep horn is characterized by comprising the following steps:
step 1, image acquisition: taking a picture by a high-speed camera in an image acquisition module to acquire an original image;
step 2, image preprocessing: filtering, enhancing and detecting the edge of the image;
step 3, the construction method of the goat horn ASM comprises the following steps: the method comprises the steps of calibrating a goat horn learning sample and carrying out ASM training;
step 4, preliminary positioning of a cavel area: carrying out initial positioning on the cavel area;
step 5, horn detection and identification: and matching the goat horn by combining an active shape model learning algorithm, matching the goat horn shape by adopting a single resolution search algorithm, and judging whether the goat horn in the initial positioning area is missing or not.
4. The image processing-based online detection method for the sheep horns of the pantograph according to claim 3, wherein the construction method for the ASM of the sheep horns in the step 3 comprises sheep horn learning sample calibration and ASM training, and specifically comprises the following steps:
(3.1) goat's horn learning sample calibration
After image preprocessing, selecting boundary points and angular points of the goat horn outline as characteristic points, and marking the goat horn characteristic points in a manual mode; in the marking process, the number of characteristic points of each goat horn image is required to be consistent and corresponding to each other, and the shape of the goat horn is described by adopting PDM (product data model), namely the shape of the goat horn image i is represented by all the characteristic points of the goat horn image:
<mrow> <msup> <mi>x</mi> <mi>i</mi> </msup> <mo>=</mo> <msup> <mrow> <mo>(</mo> <msubsup> <mi>x</mi> <mn>1</mn> <mi>i</mi> </msubsup> <mo>,</mo> <msubsup> <mi>y</mi> <mn>1</mn> <mi>i</mi> </msubsup> <mo>,</mo> <msubsup> <mi>x</mi> <mn>2</mn> <mi>i</mi> </msubsup> <mo>,</mo> <msubsup> <mi>y</mi> <mn>2</mn> <mi>i</mi> </msubsup> <mo>,</mo> <mo>...</mo> <mo>,</mo> <msubsup> <mi>x</mi> <mi>N</mi> <mi>i</mi> </msubsup> <mo>,</mo> <msubsup> <mi>y</mi> <mi>N</mi> <mi>i</mi> </msubsup> <mo>)</mo> </mrow> <mi>T</mi> </msup> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>1</mn> <mo>)</mo> </mrow> </mrow>
wherein N is the total number of characteristic points of the cavel image;
the cavel image learning sample set is represented as:
<mrow> <msup> <mi>x</mi> <mi>i</mi> </msup> <mo>=</mo> <msup> <mrow> <mo>(</mo> <msubsup> <mi>x</mi> <mn>1</mn> <mi>i</mi> </msubsup> <mo>,</mo> <msubsup> <mi>y</mi> <mn>1</mn> <mi>i</mi> </msubsup> <mo>,</mo> <msubsup> <mi>x</mi> <mn>2</mn> <mi>i</mi> </msubsup> <mo>,</mo> <msubsup> <mi>y</mi> <mn>2</mn> <mi>i</mi> </msubsup> <mo>,</mo> <mo>...</mo> <mo>,</mo> <msubsup> <mi>x</mi> <mi>N</mi> <mi>i</mi> </msubsup> <mo>,</mo> <msubsup> <mi>y</mi> <mi>N</mi> <mi>i</mi> </msubsup> <mo>)</mo> </mrow> <mi>T</mi> </msup> <mo>,</mo> <mi>i</mi> <mo>=</mo> <mn>1</mn> <mo>,</mo> <mn>2</mn> <mo>,</mo> <mn>3</mn> <mo>,</mo> <mo>...</mo> <mo>,</mo> <mi>M</mi> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>2</mn> <mo>)</mo> </mrow> </mrow>
wherein M is the total number of the horn images;
(2) ASM training
Firstly, aligning feature points, and specifically comprising the following steps:
a) the shape of a horn xiI is 1,2,3, …, M, and is translated, rotated and contracted one by onePut transform and shape x1Align to obtain a transformed shape set
b) Averaging the calculated and transformed cavel images to obtain an average shape m:
wherein,
c) translating, rotating, scaling the average shape m, andaligning;
d) will be provided withCarrying out translation, rotation and scaling transformation, and then aligning and matching with the average shape m;
e) if the average shape is converged, stopping, otherwise, turning to the step b);
the convergence in step e is determined by minimizing the sum of squares of the differences between the aligned horn shapes and the average shape, i.e., finding the transformation TiSo that the following equation is minimized:
∑|m-Ti(xi)|2(4)
the alignment of the horn images is described as: taking two horn shapes as an example, each shape has N coordinate pairs:
<mrow> <msup> <mi>x</mi> <mn>1</mn> </msup> <mo>=</mo> <msup> <mrow> <mo>(</mo> <msubsup> <mi>x</mi> <mn>1</mn> <mn>1</mn> </msubsup> <mo>,</mo> <msubsup> <mi>y</mi> <mn>1</mn> <mn>1</mn> </msubsup> <mo>,</mo> <msubsup> <mi>x</mi> <mn>2</mn> <mn>1</mn> </msubsup> <mo>,</mo> <msubsup> <mi>y</mi> <mn>2</mn> <mn>1</mn> </msubsup> <mo>,</mo> <mo>...</mo> <mo>,</mo> <msubsup> <mi>x</mi> <mi>N</mi> <mn>1</mn> </msubsup> <mo>,</mo> <msubsup> <mi>y</mi> <mi>N</mi> <mn>1</mn> </msubsup> <mo>)</mo> </mrow> <mi>T</mi> </msup> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>5</mn> <mo>)</mo> </mrow> </mrow>
<mrow> <msup> <mi>x</mi> <mn>2</mn> </msup> <mo>=</mo> <msup> <mrow> <mo>(</mo> <msubsup> <mi>x</mi> <mn>1</mn> <mn>2</mn> </msubsup> <mo>,</mo> <msubsup> <mi>y</mi> <mn>1</mn> <mn>2</mn> </msubsup> <mo>,</mo> <msubsup> <mi>x</mi> <mn>2</mn> <mn>2</mn> </msubsup> <mo>,</mo> <msubsup> <mi>y</mi> <mn>2</mn> <mn>2</mn> </msubsup> <mo>,</mo> <mo>...</mo> <mo>,</mo> <msubsup> <mi>x</mi> <mi>N</mi> <mn>2</mn> </msubsup> <mo>,</mo> <msubsup> <mi>y</mi> <mi>N</mi> <mn>2</mn> </msubsup> <mo>)</mo> </mrow> <mi>T</mi> </msup> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>6</mn> <mo>)</mo> </mrow> </mrow>
first defining a transformation momentThe matrix T, T is composed of 4 parameters, which are the rotation angle theta, the scale s and the translation vector (T)x,ty) Will shape x2And (3) carrying out transformation:
<mrow> <mi>T</mi> <mrow> <mo>(</mo> <msup> <mi>x</mi> <mn>2</mn> </msup> <mo>)</mo> </mrow> <mo>=</mo> <mfenced open = "(" close = ")"> <mtable> <mtr> <mtd> <mrow> <mi>s</mi> <mi> </mi> <mi>c</mi> <mi>o</mi> <mi>s</mi> <mi>&theta;</mi> </mrow> </mtd> <mtd> <mrow> <mo>-</mo> <mi>s</mi> <mi>i</mi> <mi>n</mi> <mi>&theta;</mi> </mrow> </mtd> </mtr> <mtr> <mtd> <mrow> <mi>s</mi> <mi> </mi> <mi>s</mi> <mi>i</mi> <mi>n</mi> <mi>&theta;</mi> </mrow> </mtd> <mtd> <mrow> <mi>s</mi> <mi> </mi> <mi>c</mi> <mi>o</mi> <mi>s</mi> <mi>&theta;</mi> </mrow> </mtd> </mtr> </mtable> </mfenced> <mfenced open = "(" close = ")"> <mtable> <mtr> <mtd> <msubsup> <mi>x</mi> <mi>i</mi> <mn>2</mn> </msubsup> </mtd> </mtr> <mtr> <mtd> <msubsup> <mi>y</mi> <mi>i</mi> <mn>2</mn> </msubsup> </mtd> </mtr> </mtable> </mfenced> <mo>+</mo> <mfenced open = "(" close = ")"> <mtable> <mtr> <mtd> <msub> <mi>t</mi> <mi>x</mi> </msub> </mtd> </mtr> <mtr> <mtd> <msub> <mi>t</mi> <mi>y</mi> </msub> </mtd> </mtr> </mtable> </mfenced> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>7</mn> <mo>)</mo> </mrow> </mrow>
is provided with
<mrow> <mi>R</mi> <mo>=</mo> <mfenced open = "(" close = ")"> <mtable> <mtr> <mtd> <mrow> <mi>s</mi> <mi> </mi> <mi>c</mi> <mi>o</mi> <mi>s</mi> <mi>&theta;</mi> </mrow> </mtd> <mtd> <mrow> <mo>-</mo> <mi>s</mi> <mi>i</mi> <mi>n</mi> <mi>&theta;</mi> </mrow> </mtd> </mtr> <mtr> <mtd> <mrow> <mi>s</mi> <mi> </mi> <mi>s</mi> <mi>i</mi> <mi>n</mi> <mi>&theta;</mi> </mrow> </mtd> <mtd> <mrow> <mi>s</mi> <mi> </mi> <mi>c</mi> <mi>o</mi> <mi>s</mi> <mi>&theta;</mi> </mrow> </mtd> </mtr> </mtable> </mfenced> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>8</mn> <mo>)</mo> </mrow> </mrow>
Transforming x with T2And x1Alignment, the optimal transformation is obtained by minimizing equation (4):
E＝[x1-Rx2-(tx,ty)T]T(9)
by calculating E pairs of unknown parameters theta, s,tx、tyAnd making a differential equation be zero, so as to solve and obtain a transformation matrix T;
second step, ASM establishment
Obtaining M training shapes after alignment processingEach shape is given by N pairs of coordinates:the average shape is set as:the covariance matrix is then:
wherein S is a 2 Nx 2N matrix;
the variation of the training shape in certain directions is obtained by the eigenvectors of the covariance matrix S, i.e. solving the linear equation:
Spk＝λkpk,k＝1,2,3,…,2N (11)
wherein, the feature vector of S is P, and P is expressed as: p ═ P (P)1,p2,…,p2N)；
For any vector X, there is a shape model parameter b, satisfying:
<mrow> <mi>x</mi> <mo>=</mo> <mover> <mi>x</mi> <mo>&OverBar;</mo> </mover> <mo>+</mo> <mi>P</mi> <mi>b</mi> <mo>=</mo> <mover> <mi>x</mi> <mo>&OverBar;</mo> </mover> <mo>+</mo> <msub> <mi>p</mi> <mn>1</mn> </msub> <msub> <mi>b</mi> <mn>1</mn> </msub> <mo>+</mo> <mo>...</mo> <mo>+</mo> <msub> <mi>p</mi> <mrow> <mn>2</mn> <mi>N</mi> </mrow> </msub> <msub> <mi>b</mi> <mrow> <mn>2</mn> <mi>N</mi> </mrow> </msub> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>12</mn> <mo>)</mo> </mrow> </mrow>
order:
<mrow> <mtable> <mtr> <mtd> <mrow> <msub> <mi>P</mi> <mi>t</mi> </msub> <mo>=</mo> <mrow> <mo>(</mo> <mrow> <msub> <mi>p</mi> <mn>1</mn> </msub> <mo>,</mo> <msub> <mi>p</mi> <mn>2</mn> </msub> <mo>,</mo> <msub> <mi>p</mi> <mn>3</mn> </msub> <mo>,</mo> <mn>...</mn> <mo>,</mo> <msub> <mi>p</mi> <mi>t</mi> </msub> </mrow> <mo>)</mo> </mrow> </mrow> </mtd> </mtr> <mtr> <mtd> <mrow> <msub> <mi>b</mi> <mi>t</mi> </msub> <mo>=</mo> <msup> <mrow> <mo>(</mo> <mrow> <msub> <mi>b</mi> <mn>1</mn> </msub> <mo>,</mo> <msub> <mi>b</mi> <mn>2</mn> </msub> <mo>,</mo> <msub> <mi>b</mi> <mn>3</mn> </msub> <mo>,</mo> <mn>...</mn> <mo>,</mo> <msub> <mi>b</mi> <mi>t</mi> </msub> </mrow> <mo>)</mo> </mrow> <mi>T</mi> </msup> </mrow> </mtd> </mtr> </mtable> <mo>,</mo> <mi>t</mi> <mo>&le;</mo> <mn>2</mn> <mi>N</mi> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>13</mn> <mo>)</mo> </mrow> </mrow>
thus, an estimate of the shape is obtained:
<mrow> <mi>x</mi> <mo>&ap;</mo> <mover> <mi>x</mi> <mo>&OverBar;</mo> </mover> <mo>+</mo> <msub> <mi>P</mi> <mi>t</mi> </msub> <msub> <mi>b</mi> <mi>t</mi> </msub> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>14</mn> <mo>)</mo> </mrow> </mrow>
<mrow> <msub> <mi>b</mi> <mi>t</mi> </msub> <mo>&ap;</mo> <msubsup> <mi>P</mi> <mi>t</mi> <mi>T</mi> </msubsup> <mrow> <mo>(</mo> <mi>x</mi> <mo>-</mo> <mover> <mi>x</mi> <mo>&OverBar;</mo> </mover> <mo>)</mo> </mrow> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>15</mn> <mo>)</mo> </mrow> </mrow>
vector btA set of variable model parameters, different b, is definedtDifferent changing shapes can be fitted;
due to biVariance and eigenvalue lambda over the training setiRelated to, biTo satisfy the following equation:
<mrow> <mo>-</mo> <mn>3</mn> <msqrt> <msub> <mi>&lambda;</mi> <mi>i</mi> </msub> </msqrt> <mo>&le;</mo> <msub> <mi>b</mi> <mi>i</mi> </msub> <mo>&le;</mo> <mn>3</mn> <msqrt> <msub> <mi>&lambda;</mi> <mi>i</mi> </msub> </msqrt> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>16</mn> <mo>)</mo> </mrow> <mo>.</mo> </mrow>
5. the image processing-based online detection method for the toe of a pantograph of claim 3, wherein the preliminary positioning of the toe area in the step 4 is as follows:
(1) the collected sheep horns of the pantograph are distributed on the left side and the right side of the image, intersection points are taken as characteristic points, and the left half-bow image and the right half-bow image respectively take the left extension line direction and the right extension line direction of the intersection points as search directions;
(2) the region at least includes two straight lines l1And l2Describing the angle of two straight lines by a straight line methodAndif they are satisfiedThe area is the area to be determined; wherein,andare respectively a straight line l1And l2The angle of inclination of (d);
after the area to be determined of the cavel is positioned, because the input image in the cavel detection algorithm is an edge image, curve data compression is directly carried out on edge line segments in the search area on the basis, and then line segments with the length smaller than a set threshold value in the edge line segments are removed; because the directions of the edge line segments of the left and right sheep horns are respectively in the ranges of 35-55 degrees and 125-145 degrees, straight lines are detected by Hough transformation, straight line segments with the angle ranges in the two ranges are searched according to results, and then the inclined edge line segment region is used as an initial positioning region of the sheep horn.
6. The image processing-based online detection method for the sheep horns of the pantograph according to claim 3, wherein the sheep horn detection and identification in the step 5 are as follows:
(1) initializing the goat's horn shape from the average shape in the goat's horn active shape model and the initial position of the goat's horn, as follows:
(2) searching along the boundary normal direction at each mark point of the initial positioning area of the cavel, further obtaining a pixel point with the maximum gradient, marking the point as an optimal target point, moving the mark point to the optimal target point, and if a new target point is not searched, the position of the mark point does not move;
(3) after the mark point is moved, the shape is changed, and a displacement vector exists between the changed shape and the initialized horn shapeAs can be seen from equation (15), the expression after displacement is:
<mrow> <mover> <mi>x</mi> <mo>&OverBar;</mo> </mover> <mo>+</mo> <mi>&zeta;</mi> <mover> <mi>x</mi> <mo>&OverBar;</mo> </mover> <mo>&ap;</mo> <mover> <mi>x</mi> <mo>&OverBar;</mo> </mover> <mo>+</mo> <msub> <mi>P</mi> <mi>t</mi> </msub> <mrow> <mo>(</mo> <mrow> <msub> <mi>b</mi> <mi>t</mi> </msub> <mo>+</mo> <msub> <mi>&zeta;b</mi> <mi>t</mi> </msub> </mrow> <mo>)</mo> </mrow> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>18</mn> <mo>)</mo> </mrow> </mrow>
derived from equations (15) and (18):
<mrow> <msub> <mi>&zeta;b</mi> <mi>t</mi> </msub> <mo>&ap;</mo> <msubsup> <mi>P</mi> <mi>t</mi> <mi>T</mi> </msubsup> <mi>&zeta;</mi> <mover> <mi>x</mi> <mo>&OverBar;</mo> </mover> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>19</mn> <mo>)</mo> </mrow> </mrow>
(4) repeating the steps (2) to (3), and if the preset times of repetition are carried out, p2N,p2N-1…, if the value is less than the threshold value sigma and sigma tends to zero, judging that the goat horn exists, otherwise, judging that the goat horn in the image is missing.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710719718.XA CN107590441A (en) | 2017-08-21 | 2017-08-21 | A kind of pantograph goat's horn on-line measuring device and method based on image procossing |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710719718.XA CN107590441A (en) | 2017-08-21 | 2017-08-21 | A kind of pantograph goat's horn on-line measuring device and method based on image procossing |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107590441A true CN107590441A (en) | 2018-01-16 |
Family
ID=61042453
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710719718.XA Pending CN107590441A (en) | 2017-08-21 | 2017-08-21 | A kind of pantograph goat's horn on-line measuring device and method based on image procossing |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107590441A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111260629A (en) * | 2020-01-16 | 2020-06-09 | 成都地铁运营有限公司 | Pantograph structure abnormity detection algorithm based on image processing |
| CN112132789A (en) * | 2020-08-30 | 2020-12-25 | 南京理工大学 | Pantograph online detection device and method based on cascade neural network |
-
2017
- 2017-08-21 CN CN201710719718.XA patent/CN107590441A/en active Pending
Non-Patent Citations (1)
| Title |
|---|
| 陈双: "基于图像处理的受电弓故障检测算法研究", 《中国优秀硕士学位论文全文数据库 工程科技Ⅱ辑》 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111260629A (en) * | 2020-01-16 | 2020-06-09 | 成都地铁运营有限公司 | Pantograph structure abnormity detection algorithm based on image processing |
| CN112132789A (en) * | 2020-08-30 | 2020-12-25 | 南京理工大学 | Pantograph online detection device and method based on cascade neural network |
| CN112132789B (en) * | 2020-08-30 | 2022-10-25 | 南京理工大学 | Pantograph online detection device and method based on cascade neural network |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110569704B (en) | Multi-strategy self-adaptive lane line detection method based on stereoscopic vision | |
| CN104484648B (en) | Robot variable visual angle obstacle detection method based on outline identification | |
| CN105678689B (en) | High-precision map data registration relation determining method and device | |
| CN111223133B (en) | Registration method of heterogeneous images | |
| CN101839692B (en) | Method for measuring three-dimensional position and stance of object with single camera | |
| CN103559711B (en) | Based on the method for estimating of three dimensional vision system characteristics of image and three-dimensional information | |
| US11948344B2 (en) | Method, system, medium, equipment and terminal for inland vessel identification and depth estimation for smart maritime | |
| CN105674880B (en) | Contact net geometric parameter measurement method and system based on binocular principle | |
| CN106996748A (en) | Wheel diameter measuring method based on binocular vision | |
| CN104121902B (en) | Implementation method of indoor robot visual odometer based on Xtion camera | |
| CN106023211B (en) | Robot graphics' localization method and system based on deep learning | |
| CN104809738B (en) | A kind of air bag overall size detection method based on binocular vision | |
| CN106355147A (en) | Acquiring method and detecting method of live face head pose detection regression apparatus | |
| CN107014294A (en) | A kind of contact net geometric parameter detection method and system based on infrared image | |
| CN101398886A (en) | Rapid three-dimensional face identification method based on bi-eye passiveness stereo vision | |
| CN107016353B (en) | A kind of integrated method and system of variable resolution target detection and identification | |
| CN102609684A (en) | Human body posture detection method and device | |
| CN111402330A (en) | Laser line key point extraction method based on plane target | |
| CN111524174A (en) | Binocular vision three-dimensional construction method for moving target of moving platform | |
| CN114820474A (en) | Train wheel defect detection method based on three-dimensional information | |
| CN116152068A (en) | Splicing method for solar panel images | |
| CN107590441A (en) | A kind of pantograph goat's horn on-line measuring device and method based on image procossing | |
| CN106408589A (en) | Vehicle-mounted overlooking camera based vehicle movement measurement method | |
| CN114066954B (en) | Feature extraction and registration method for multi-modal image | |
| CN110533971A (en) | A kind of intelligent tutoring system deeply interacted |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20180116 |