CN115393195B - Generator stator core infrared image panorama stitching method based on visible light assistance - Google Patents

Generator stator core infrared image panorama stitching method based on visible light assistance Download PDF

Info

Publication number
CN115393195B
CN115393195B CN202211152638.8A CN202211152638A CN115393195B CN 115393195 B CN115393195 B CN 115393195B CN 202211152638 A CN202211152638 A CN 202211152638A CN 115393195 B CN115393195 B CN 115393195B
Authority
CN
China
Prior art keywords
image
visible light
iron core
infrared
corrected
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN202211152638.8A
Other languages
Chinese (zh)
Other versions
CN115393195A (en
Inventor
郭钰静
谭鋆
冉应兵
关苏敏
曹慧明
徐铬
陈练
郭筱飞
冯维颖
范平安
杨辉
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
China Yangtze Power Co Ltd
Original Assignee
China Yangtze Power Co Ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by China Yangtze Power Co Ltd filed Critical China Yangtze Power Co Ltd
Priority to CN202211152638.8A priority Critical patent/CN115393195B/en
Publication of CN115393195A publication Critical patent/CN115393195A/en
Application granted granted Critical
Publication of CN115393195B publication Critical patent/CN115393195B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

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
    • G06T3/40—Scaling of whole images or parts thereof, e.g. expanding or contracting
    • G06T3/4038—Image mosaicing, e.g. composing plane images from plane sub-images
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T5/00—Image enhancement or restoration
    • G06T5/80—Geometric correction
    • 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
    • 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
    • G06T2207/10—Image acquisition modality
    • G06T2207/10048—Infrared image

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Image Processing (AREA)

Abstract

A generator stator core infrared image panorama stitching method based on visible light assistance comprises the following steps: distortion correction of a single camera, and binocular calibration of a visible light camera and an infrared camera; a cradle head carrying a visible light camera and an infrared camera is adopted to rotate for one circle at the center of a stator, and a circle of visible light image and infrared image of a stator core of the generator are collected; establishing a color label area model and an iron core tooth slot part model according to a visible light image of a stator iron core; carrying out coordinate transformation on the infrared image, and aligning the infrared image with the visible light image to obtain an aligned infrared image; correcting the aligned infrared image to obtain an accurately corrected infrared image; and splicing the precisely corrected infrared images into a panoramic image. The method can effectively realize panoramic stitching of the stator core infrared images, is convenient for analyzing the distribution condition of the core temperature in the stator core magnetization test process, locates the fault point position in real time, and has high application value.

Description

Generator stator core infrared image panorama stitching method based on visible light assistance
Technical Field
The invention relates to the technical field of stator core temperature detection, in particular to a generator stator core infrared image panorama stitching method based on visible light assistance.
Background
The stator core magnetization test is carried out after the generator set is newly installed and the silicon steel sheets are reassembled or replaced and repaired, in the test process, a tester needs to detect the temperature of the stator core of the generator set, check whether the temperature rise of each part exceeds a specified value, and timely discover local overheating phenomena caused by potential safety hazards such as loosening among the silicon steel sheets, short circuit among the silicon steel sheets and the like, so that the safe and stable operation of the generator after operation is ensured. The current stator core temperature detection method is that 2 testers hold an infrared imager to finish stator core temperature detection in a machine pit, manual detection cannot count the temperature distribution of the whole stator core, and the defects can be solved by panoramic stitching of infrared images of the stator core of the generator based on the assistance of visible light.
The panoramic stitching of the infrared images of the stator iron core of the generator based on the assistance of visible light is realized by calculating correction parameters and stitching parameters according to the visible light images and then stitching the infrared images. Currently, popular image stitching algorithms include a stitching algorithm based on SIFT feature matching, a stitching algorithm based on SURF feature matching, an image stitching algorithm based on feature classification, and the like, but most of the image stitching algorithms are not applicable to the situation that more features exist in an image, and the situation that the stator core is symmetrical and the features are single.
Disclosure of Invention
In order to solve the technical problems, the invention provides the panoramic stitching method of the infrared images of the stator iron core of the generator based on the visible light assistance, which can effectively realize panoramic stitching of the infrared images of the stator iron core, is convenient for analyzing the distribution condition of the iron core temperature in the magnetizing test process of the stator iron core, locates the fault point position in real time, has high stitching quality and good robustness, and has very high application value.
The technical scheme adopted by the invention is as follows:
a generator stator core infrared image panorama stitching method based on visible light assistance comprises the following steps:
step 1: monocular camera calibration is carried out on the visible light camera and the infrared camera respectively, so that visible light image distortion correction parameters and infrared image distortion correction parameters are obtained; performing binocular camera calibration on the two cameras to obtain a coordinate mapping relation between a visible light image and an infrared image;
step 2: collecting visible light images and infrared images of a stator core of the generator;
step 3: selecting the step 2 to obtain a visible light image, and performing experimental modeling on the visible light image;
step 4: correcting the visible light image and the infrared image of the stator core according to the distortion correction parameters obtained in the step 1 to obtain a preliminarily corrected visible light image and a preliminarily corrected infrared image; according to the coordinate mapping relation between the visible light image and the infrared image obtained in the step 1, carrying out coordinate transformation on the preliminarily corrected infrared image, aligning the preliminarily corrected infrared image with the preliminarily corrected visible light image coordinate, and obtaining an aligned infrared image;
step 5: according to the initially corrected visible light image obtained in the step 4, iron core correction parameters are obtained through calculation, and according to the iron core correction parameters, the aligned infrared image obtained in the step 4 is corrected, so that an accurately corrected infrared image is obtained;
step 6: and (5) splicing the accurately corrected infrared images obtained in the step (5) into a panoramic image.
In the step 1, monocular camera calibration is performed on a visible light camera and an infrared camera respectively to obtain a visible light image distortion correction parameter and an infrared image distortion correction parameter; and (3) performing binocular camera calibration on the visible light camera and the infrared camera to obtain a coordinate mapping relation between the visible light image and the infrared image.
In the step 2, a cradle head carrying a visible light camera and an infrared camera is adopted to rotate around the center of a circle of a stator, and a circle of visible light image and infrared image of a stator core of the generator are collected.
In the step 3, a visible light image obtained in the step 2 is selected, a rectangular frame is adopted to respectively frame and select a color label area and a single tooth area in the visible light image, a color label template and a rectangle cRect, a tooth template and a rectangle trigger are obtained, and the number Rows and the number Cols of the teeth of the iron core are recorded and stored.
In the step 5, the iron core correction parameters include pixel intervals of adjacent teeth of the iron core, calculating the radius of the iron core pixels and calculating the center coordinates of the iron core image.
In the step 5, the method for calculating the correction parameters of the iron core according to the preliminarily corrected visible light image comprises the following steps:
s5.1, graying treatment: carrying out graying treatment on the preliminarily corrected visible light image to obtain a gray image GrayI;
s5.2, edge processing: respectively carrying out X-direction and Y-direction edge processing on the gray level image GrayI to obtain an iron core edge image edge I x And core edge image edge y The method comprises the steps of carrying out a first treatment on the surface of the The formula for edge processing is as follows:
where x1, x2, y1, y2 are convolution kernels, which represent convolutions.
S5.3, binarization processing: respectively for iron core edge images edge I x And edge I y Performing self-adaptive threshold processing to obtain an iron core binarization image BinaryX and a binarization image BinaryY;
s5.4, calculating the pixel interval Dis between adjacent teeth of the iron core in the horizontal direction x And a vertical direction pixel pitch Dis y : projecting the iron core binarized image BinaryX to the horizontal direction to obtain a horizontal projection histogram, and calculating to obtain Dis according to the average distance between adjacent extreme points of the horizontal projection histogram x Projecting the iron core binarized image BinaryY to the vertical direction to obtain a vertical projection histogram, and calculating to obtain Dis according to the average distance between adjacent extreme points of the vertical projection histogram y ；
S5.5, calculating the radius R of the iron core pixel: according to the horizontal pixel interval Dis of adjacent teeth of the iron core x And the number of columns of the iron core Cols, calculating to obtain the radius R of the iron core pixel, wherein the calculation formula is as follows:
where pi is the circumference ratio.
S5.6, calculating the center coordinates (x, y) of the iron core image: projecting a binarized image BinaryY in a vertical direction to obtain a vertical projection histogram, taking the maximum value position in the vertical projection histogram as the y coordinate of the center of the iron core image, wherein the x coordinate of the center of the iron core image is 1/(2 x width), and the method comprises the following steps: width is the pixel width of the image.
In the step 5, the aligned infrared images are corrected according to the correction parameters of the iron core, and the formula is as follows:
m 1 ＝x 0 -x
n 1 ＝y 0 -y
x 1 ＝round(θ×R+x)
y 1 ＝round(n 1 ×cosθ+y)
I 1 (x 1 ，y 1 )＝I 0 (x 0 ，y 0 )
wherein the round function represents a decimal point rounding (x) 0 ，y 0 ) For aligned infrared images I 0 Pixel coordinates of (x) 1 ，y 1 ) Infrared image I for accurate correction 1 Pixel coordinates, m 1 、n 1 θ is an intermediate variable in the calculation process. In the step 6, the step of stitching the corrected infrared images into a panoramic image is as follows:
s6.1: according to the pixel spacing Dis of adjacent teeth of the iron core in the horizontal direction x And a vertical direction pixel pitch Dis y The number Rows and the number Cols of Rows of the teeth of the iron core automatically generate an iron core virtual panorama, and the height H and the width W of the virtual panorama are respectively as follows:
W＝Dis x ×Cols；
H＝Dis y ×Rows；
s6.2: according to the row number and the column number of the iron core teeth, all the teeth coordinates in the virtual panorama are calculated, a teeth coordinate lookup table is generated, and the teeth coordinates (x, y) are calculated by the following steps:
x＝col×Dis x
y＝row×Dis y
wherein row and col are the tooth row number and column number, respectively.
S6.3: according to the iron core correction parameters in the step 5, accurately correcting the initially corrected visible light image obtained in the step 4 to obtain an accurately corrected visible light image;
s6.4: using a color label template as a template, adopting template matching to the accurately corrected visible light image to obtain a color label positioning position, identifying the positioned color label to obtain a label identification result, and obtaining the line number and the column number of the iron core where the label is positioned according to the label identification result;
s6.5: according to the edge processing method described in S5.2, the precisely corrected visible light image and the tooth template are processed to obtain a corrected iron core contour image and a tooth contour template, the tooth contour template is used as a template, template matching is adopted to the corrected iron core contour image to obtain a set composed of the positioned iron core tooth coordinates, and the set is recorded as
S6.7: cyclic access tooth setAccording to the coordinates of each element, combining the positioning position of the tag, the line number and the column number of the iron core where the tag is positioned, and obtaining the line number and the column number of the element, the calculating method comprises the following steps:
wherein, (x) l ，y l ) For locating the tag, row l 、col l The row number and the column number of the iron core where the label is located are respectively (x) i ，y i ) Is the coordinates of the element, r i 、c i The row number and column number of the element, respectively.
The line number and the column number of the elements are used as input, the coordinates in the virtual panorama corresponding to the elements are obtained according to the lookup table, the coordinates in the virtual panorama corresponding to all the elements form a new set, and the new set is recorded asAnd->One-to-one correspondence;
s6.8: computing a set using the SANSAC methodTo the collection->The infrared image corrected accurately is spliced into the virtual panorama through the optimal single mapping matrix;
s6.9: and (4) circularly executing the steps (4), 5 and S6.1-S6.8), and splicing the infrared images of the stator core of the generator in a circle into the virtual panorama to finish panorama splicing of the infrared images.
The invention discloses a visible light-assisted panoramic stitching method for infrared images of a stator core of a generator, which has the following technical effects:
1) The method can effectively realize panoramic splicing of the infrared images of the stator core of the generator, is convenient for analyzing the panoramic temperature distribution of the stator core in the magnetizing test process of the stator core, and has high splicing quality and good robustness.
2) The method adopts visible light to assist in calculation, and solves the problems of low resolution of infrared images and low feature calculation precision.
3) The method adopts a unique image accurate correction method to realize asymmetric distribution distortion correction of the visible light image and the infrared image.
4) The method adopts the unique thought of the virtual panorama and solves the problem of image splicing of the symmetrical structure of the stator core of the generator.
Drawings
FIG. 1 is a flow chart of the present invention.
Fig. 2 is a visible light diagram of a color-labeled generator stator core.
Fig. 3 is an infrared image of a stator core of a generator.
Fig. 4 is a graph of experimental modeling effects.
Fig. 5 is a primarily corrected visible light image.
Fig. 6 is an infrared image of a preliminary correction.
Fig. 7 is an aligned infrared image.
Fig. 8 is a virtual panorama.
Fig. 9 is a precisely rectified visible light image.
FIG. 10 color tag matching schematic
Fig. 11 corrected core profile image
Fig. 12 is a stitched infrared panoramic image.
Detailed Description
As shown in fig. 1, the panorama stitching method for the infrared image of the stator core of the generator based on the assistance of visible light calculates correction parameters and stitching parameters by using the visible light image, and realizes panorama stitching on the infrared image, and the specific implementation steps are as follows:
(1) Calibrating a camera:
monocular camera calibration is carried out on the visible light camera and the infrared camera respectively, so that visible light image distortion correction parameters and infrared image distortion correction parameters are obtained; performing binocular camera calibration on the two cameras to obtain a coordinate mapping relation between a visible light image and an infrared image;
monocular camera calibration: and (3) adopting a Zhang Zhengyou camera calibration method to respectively carry out distortion correction on the visible light camera and the infrared camera to obtain internal parameters, external parameters and distortion coefficients of the two cameras, namely, the visible light image distortion correction parameters and the infrared image distortion correction parameters.
Calibrating a binocular camera: and respectively shooting visible light images and infrared image groups with different multiplying powers at different distances, and performing feature matching on each group of images to obtain a mapping matrix of each group of visible light images and infrared images. The mapping matrix H, the distance b and the multiplying power a form a mapping table, as shown in table 1,
table 1 binocular camera calibration parameter map
The mapping table 1 is the coordinate mapping relation between the visible light image and the infrared image. And inquiring in the mapping table according to the distance and the multiplying power to obtain a mapping matrix. The formula for calculating the mapping matrix through the matching point pairs of the visible light image and the infrared image is as follows:
X 1 ＝HX 0
wherein X is 0 X is the set of the matching points of the infrared image 1 And H is a mapping matrix, wherein the mapping matrix is a set of visible light image matching points.
(2) And (3) image acquisition:
color labels are sequentially attached to the inner teeth of the stator iron core before collection, the row number and the column number of the teeth corresponding to each color label are recorded, the color labels are strictly aligned with the iron core teeth, the upper left corner of a color label square frame is aligned with the upper left corner of the iron core tooth square frame, and the alignment effect of the color labels and the iron core teeth is shown in figure 2.
The cradle head carrying the visible light camera and the infrared camera rotates around the center of the stator for one circle, and a circle of visible light image and infrared image of the stator core of the generator are collected. As shown in fig. 2, a visible light image of the stator core; fig. 3 shows an infrared image of the stator core.
(3) And (3) experimental modeling:
and selecting a visible light image, respectively selecting a color label area and a single tooth area in the visible light image by adopting a rectangular frame through modeling software, obtaining a color label template, a rectangular cRect, a tooth template and a rectangular direction, and recording and storing the number Rows and the number Cols of the teeth of the iron core.
As shown in fig. 4, the red area is a frame-selected color label area and a single tooth area, and the frame-selected area is proposed to obtain a color label area sub-image and a tooth area sub-image, wherein the two sub-images are a color label template and a tooth template, and the positions and the sizes of the two areas in the visible light image are a color label rectangle cRect and a tooth rectangle rect.
(4) Alignment of coordinates:
and correcting the visible light image and the infrared image of the stator core according to the distortion correction parameters respectively to obtain a preliminarily corrected visible light image shown in fig. 5 and a preliminarily corrected infrared image shown in fig. 6.
And according to the coordinate mapping relation between the visible light image and the infrared image, performing perspective transformation on the primarily corrected infrared image, and aligning the primarily corrected infrared image with the primarily corrected visible light image coordinate to obtain an aligned infrared image, as shown in fig. 7.
(5) Image correction:
according to the preliminarily corrected visible light image, calculating to obtain iron core correction parameters, wherein the iron core correction parameters comprise pixel intervals of adjacent teeth of an iron core, calculating the pixel radius of the iron core and calculating the center coordinates of the iron core image, and correcting the aligned infrared image according to the iron core correction parameters to obtain an accurately corrected infrared image;
1) The method for calculating the correction parameters of the iron core is as follows:
A. graying treatment; carrying out graying treatment on the preliminarily corrected visible light image to obtain a gray image GrayI;
B. edge processing; respectively carrying out X-direction and Y-direction edge processing on the gray level image to obtain an iron core edge image edge I x And core edge image edge y The method comprises the steps of carrying out a first treatment on the surface of the The formula for edge processing is as follows:
where x1, x2, y1, y2 are convolution kernels, which represent convolutions.
C. Binarization processing; respectively for iron core edge images edge I x And edge I y Performing self-adaptive threshold processing to obtain an iron core binarization image BinaryX and a binarization image BinaryY;
D. calculating the horizontal pixel distance Dis of adjacent teeth of the iron core x And a vertical direction pixel pitch Dis y The method comprises the steps of carrying out a first treatment on the surface of the Projecting the iron core binarized image BinaryX to the horizontal direction to obtain a horizontal projection histogram, and calculating to obtain Dis according to the average distance between adjacent extreme points of the horizontal projection histogram x Projecting the iron core binarized image BinaryY to the vertical direction to obtain a vertical projection histogram, and calculating to obtain Dis according to the average distance between adjacent extreme points of the vertical projection histogram y The method comprises the steps of carrying out a first treatment on the surface of the The horizontal projection histogram and the vertical projection histogram are discrete functions, respectively:
H x ＝m i ，0≤i＜col
H y ＝n j ，0≤j＜row
wherein H is x For horizontal projection histogram, H y For vertical projection histogram, col, row are respectively the pixel width of the iron core binarized image BinaryX and the pixel height of the iron core binarized image BinaryY, m i N is the number of pixels with gray value of 255 in the vertical direction of the ith pixel width position j Is the number of pixels with a gray value of 255 in the horizontal direction of the j-th pixel height position.
E. Calculating the radius R of the iron core pixel; according to the horizontal pixel interval Dis of adjacent teeth of the iron core x And the number of columns of the iron core Cols, calculating to obtain the radius R of the iron core pixel, wherein the calculation formula is as follows:
where pi is the circumference ratio.
F. Calculating the center coordinates (x, y) of the iron core image; projecting a binarized image BinaryY in a vertical direction to obtain a vertical projection histogram, taking the maximum value position in the histogram as the y coordinate of the center of the iron core image, wherein the x coordinate of the center of the iron core image is 1/(2 x width), and the method comprises the following steps: width is the pixel width of the image.
2) The infrared images aligned according to the iron core correction parameters are corrected according to the following formula:
m 1 ＝x 0 -x
n 1 ＝y 0 -y
x 1 ＝round(θ×R+x)
y 1 ＝round(n 1 ×cosθ+y)
I 1 (x 1 ，y 1 )＝I 0 (x 0 ，y 0 )
wherein the round function represents a decimal point rounding (x) 0 ，y 0 ) For aligned infrared images I 0 Pixel coordinates of (x) 1 ，y 1 ) Infrared image I for accurate correction 1 Pixel coordinates, m 1 、n 1 θ is an intermediate variable in the calculation process.
(6) Panoramic stitching:
the corrected infrared images are spliced into a panoramic image, and the steps of splicing the infrared images into the panoramic image are as follows:
a. according to the pixel spacing Dis of adjacent teeth of the iron core in the horizontal direction x And a vertical direction pixel pitch Dis y The virtual panorama of the iron core is automatically generated according to the number Rows and the number Cols of Rows of the iron core teeth, and is automatically generated according to the inherent characteristics of the paired structures of the iron core teeth, and is not displayed in actual calculation as shown in a visual mode of the virtual panorama in FIG. 8. The height H and width W of the virtual panorama are respectively:
W＝Dis x ×Cols；
H＝Dis y ×Rows；
b. according to the row number and the column number of the iron core teeth, all the teeth coordinates in the virtual panorama are calculated, and the row number, the column number and the teeth coordinates form a teeth coordinate lookup table, as shown in table 2,
table 2 tooth coordinates lookup table
The tooth part coordinates can be found in the lookup table by searching the row number and the column number, and the tooth part coordinates (x, y) calculating method comprises the following steps:
x＝col×Dis x
y＝row×Dis y
wherein row and col are the tooth row number and column number, respectively.
c. And (3) accurately correcting the initially corrected visible light image according to the iron core correction parameters in the step (5), and obtaining an accurately corrected visible light image, as shown in fig. 9.
d. Using a color label template as a template, matching the precisely corrected visible light image by using the template to obtain a color label positioning position, wherein the matching effect is as shown in fig. 10, identifying the positioned color label to obtain an identification result, and obtaining the line number and the column number of the iron core where the label is positioned according to the label identification result;
e. and B, processing the precisely corrected visible light image and the tooth template according to the edge processing method in the step B to obtain a corrected iron core contour image and a corrected tooth contour template, as shown in fig. 11. The tooth profile template is used as a template, template matching is adopted to the corrected iron core profile image, and a set formed by the positioned iron core tooth coordinates can be obtained and recorded as
f. Cyclic access tooth setAccording to the coordinates of each element, combining the positioning position of the tag, the line number and the column number of the iron core where the tag is positioned, and obtaining the line number and the column number of the element, the calculating method comprises the following steps:
wherein, (x) l ，y l ) For locating the tag, row l 、col l The row number and the column number of the iron core where the label is located are respectively (x) i ，y i ) Is the coordinates of the element, r i 、c i The row number and column number of the element, respectively.
The line number and the column number of the elements are used as input, the coordinates in the virtual panorama corresponding to the elements can be obtained according to the lookup table, the coordinates in the virtual panorama corresponding to all the elements form a new set, and the new set is recorded asAnd->One-to-one correspondence;
g. computing sets using SANSAC methodTo the collection->Is a projection matrix of 3*3, by means of which the set +.>Is projected to the set +.>In that the precisely rectified infrared image can be stitched into the virtual panorama by means of an optimal single mapping matrix, as shown in fig. 12.
h. And (5) circularly executing the steps (4) and (5) and the steps a-g, and splicing the infrared images of the stator iron cores of the generator into the virtual panorama to finish panorama splicing of the infrared images.

Claims (8)

1. The generator stator core infrared image panorama stitching method based on the visible light assistance is characterized by comprising the following steps of:
step 1: calibrating a visible light camera and an infrared camera respectively to obtain a visible light image distortion correction parameter, an infrared image distortion correction parameter and a coordinate mapping relation of a visible light image and an infrared image;
step 2: collecting visible light images and infrared images of a stator core of the generator;
step 3: selecting the step 2 to obtain a visible light image, and performing experimental modeling on the visible light image;
step 4: correcting the visible light image and the infrared image of the stator core according to the distortion correction parameters obtained in the step 1 to obtain a preliminarily corrected visible light image and a preliminarily corrected infrared image; according to the coordinate mapping relation between the visible light image and the infrared image obtained in the step 1, carrying out coordinate transformation on the preliminarily corrected infrared image, aligning the preliminarily corrected infrared image with the preliminarily corrected visible light image coordinate, and obtaining an aligned infrared image;
step 5: according to the initially corrected visible light image obtained in the step 4, iron core correction parameters are obtained through calculation, and according to the iron core correction parameters, the aligned infrared image obtained in the step 4 is corrected, so that an accurately corrected infrared image is obtained;
step 6: and (5) splicing the accurately corrected infrared images obtained in the step (5) into a panoramic image.
2. The visible light-assisted generator stator core infrared image panorama stitching method according to claim 1, wherein the method comprises the following steps of: in the step 1, monocular camera calibration is performed on a visible light camera and an infrared camera respectively to obtain a visible light image distortion correction parameter and an infrared image distortion correction parameter; and (3) performing binocular camera calibration on the visible light camera and the infrared camera to obtain a coordinate mapping relation between the visible light image and the infrared image.
3. The visible light-assisted generator stator core infrared image panorama stitching method according to claim 1, wherein the method comprises the following steps of: in the step 2, a cradle head carrying a visible light camera and an infrared camera is adopted to rotate around the center of a circle of a stator, and a circle of visible light image and infrared image of a stator core of the generator are collected.
4. The visible light-assisted generator stator core infrared image panorama stitching method according to claim 1, wherein the method comprises the following steps of: in the step 3, a visible light image obtained in the step 2 is selected, a rectangular frame is adopted to respectively frame and select a color label area and a single tooth area in the visible light image, a color label template and a rectangle cRect, a tooth template and a rectangle trigger are obtained, and the number Rows and the number Cols of the teeth of the iron core are recorded and stored.
5. The visible light-assisted generator stator core infrared image panorama stitching method according to claim 1, wherein the method comprises the following steps of: in the step 5, the iron core correction parameters include pixel intervals of adjacent teeth of the iron core, calculating the radius of the iron core pixels and calculating the center coordinates of the iron core image.
6. The visible light-assisted generator stator core infrared image panorama stitching method according to claim 1, wherein the method comprises the following steps of: in the step 5, the method for calculating the correction parameters of the iron core according to the preliminarily corrected visible light image comprises the following steps:
s5.1, graying treatment: carrying out graying treatment on the preliminarily corrected visible light image to obtain a gray image GrayI;
s5.2, edge processing: respectively carrying out X-direction and Y-direction edge processing on the gray level image GrayI to obtain an iron core edge image edge I x And core edge image edge y The method comprises the steps of carrying out a first treatment on the surface of the The formula for edge processing is as follows:
wherein x1, x2, y1, y2 are convolution kernels, which represent convolutions;
s5.3, binarization processing: respectively for iron core edge images edge I x And edge I y Performing self-adaptive threshold processing to obtain an iron core binarization image BinaryX and a binarization image BinaryY;
s5.4, calculating the pixel interval Dis between adjacent teeth of the iron core in the horizontal direction x And a vertical direction pixel pitch Dis y : projecting the iron core binarized image BinaryX to the horizontal direction to obtain a horizontal projection histogram, and calculating to obtain Dis according to the average distance between adjacent extreme points of the horizontal projection histogram x Projecting the iron core binarized image BinaryY to the vertical direction to obtain a vertical projection histogram, and calculating to obtain Dis according to the average distance between adjacent extreme points of the vertical projection histogram y ；
S5.5, calculating the radius R of the iron core pixel: according to the horizontal pixel interval Dis of adjacent teeth of the iron core x And the number of columns of the iron core Cols, calculating to obtain the radius R of the iron core pixel, wherein the calculation formula is as follows:
s5.6, calculating the center coordinates (x, y) of the iron core image: projecting a binarized image BinaryY in a vertical direction to obtain a vertical projection histogram, taking the maximum value position in the vertical projection histogram as the y coordinate of the center of the iron core image, wherein the x coordinate of the center of the iron core image is 1/(2 x width), and the method comprises the following steps: width is the pixel width of the image.
7. The visible light-assisted generator stator core infrared image panorama stitching method according to claim 1 or 6, wherein the method comprises the following steps of: in the step 5, the aligned infrared images are corrected according to the correction parameters of the iron core, and the formula is as follows:
m 1 ＝x 0 -x
n 1 ＝y 0 -y
x 1 ＝round(θ×R+x)
y 1 ＝round(n 1 ×cosθ+y)
I 1 (x 1 ，y 1 )＝I 0 (x 0 ，y 0 )
wherein the round function represents a decimal point rounding (x) 0 ，y 0 ) For aligned infrared images I 0 Pixel coordinates of (x) 1 ，y 1 ) Infrared image I for accurate correction 1 Pixel coordinates, m 1 、n 1 θ is an intermediate variable in the calculation process.
8. The visible light assisted generator stator core infrared image panorama stitching method according to claim 6, wherein the method comprises the following steps: in the step 6, the corrected infrared images are spliced into a panoramic image, which comprises the following steps:
s6.1: according to the pixel spacing Dis of adjacent teeth of the iron core in the horizontal direction x And a vertical direction pixel pitch Dis y Automatic generation of virtual full iron core by row Rows and column Cols of iron core teethThe height H and width W of the scenery map and the virtual panorama are respectively:
W＝Dis x ×Cols；
H＝Dis y ×Rows；
s6.2: according to the row number and the column number of the iron core teeth, all the teeth coordinates in the virtual panorama are calculated, a teeth coordinate lookup table is generated, and the teeth coordinates (x, y) are calculated by the following steps:
x＝col×Dis x
y＝row×Dis y
wherein row and col are tooth row number and column number respectively;
s6.3: according to the iron core correction parameters in the step 5, accurately correcting the initially corrected visible light image obtained in the step 4 to obtain an accurately corrected visible light image;
s6.4: using a color label template as a template, adopting template matching to the accurately corrected visible light image to obtain a color label positioning position, identifying the positioned color label to obtain a label identification result, and obtaining the row number and the column number of the iron core where the label is positioned according to the label identification result;
s6.5: according to the edge processing method described in S5.2, the precisely corrected visible light image and the tooth template are processed to obtain a corrected iron core contour image and a tooth contour template, the tooth contour template is used as a template, template matching is adopted to the corrected iron core contour image to obtain a set composed of the positioned iron core tooth coordinates, and the set is recorded as
S6.7: cyclic access tooth setAccording to the coordinates of each element, combining the positioning position of the tag, the line number and the column number of the iron core where the tag is positioned, and obtaining the line number and the column number of the element, the calculating method comprises the following steps:
wherein, (x) l ，y l ) For locating the tag, row l 、col l The row number and the column number of the iron core where the label is located are respectively (x) i ，y i ) Is the coordinates of the element, r i 、c i The row number and the column number of the element respectively;
according to the obtained coordinates in the virtual panorama corresponding to the elements, the coordinates in the virtual panorama corresponding to all the elements are formed into a new set, and recorded asAnd->One-to-one correspondence;
s6.8: computing a set using the SANSAC methodTo the collection->The infrared image corrected accurately is spliced into the virtual panorama through the optimal single mapping matrix;
s6.9: and (4) circularly executing the steps (4), 5 and S6.1-S6.8), and splicing the infrared images of the stator core of the generator in a circle into the virtual panorama to finish panorama splicing of the infrared images.
CN202211152638.8A 2022-09-21 2022-09-21 Generator stator core infrared image panorama stitching method based on visible light assistance Active CN115393195B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202211152638.8A CN115393195B (en) 2022-09-21 2022-09-21 Generator stator core infrared image panorama stitching method based on visible light assistance

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202211152638.8A CN115393195B (en) 2022-09-21 2022-09-21 Generator stator core infrared image panorama stitching method based on visible light assistance

Publications (2)

Publication Number Publication Date
CN115393195A CN115393195A (en) 2022-11-25
CN115393195B true CN115393195B (en) 2023-11-28

Family

ID=84126281

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202211152638.8A Active CN115393195B (en) 2022-09-21 2022-09-21 Generator stator core infrared image panorama stitching method based on visible light assistance

Country Status (1)

Country Link
CN (1) CN115393195B (en)

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20100045809A1 (en) * 2008-08-22 2010-02-25 Fluke Corporation Infrared and visible-light image registration
CN109146930A (en) * 2018-09-20 2019-01-04 河海大学常州校区 A kind of electric power calculator room equipment is infrared and visible light image registration method

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20100045809A1 (en) * 2008-08-22 2010-02-25 Fluke Corporation Infrared and visible-light image registration
CN109146930A (en) * 2018-09-20 2019-01-04 河海大学常州校区 A kind of electric power calculator room equipment is infrared and visible light image registration method

Non-Patent Citations (4)

* Cited by examiner, † Cited by third party
Title
"基于R-MI-rényi测度的可见光与红外图像配准";陈震 等;《电子测量与仪器学报》;第32卷(第1期);全文 *
"基于双相机的激光照射性能监测系统图像畸变矫正方法";徐向锴 等;《长春理工大学学报（自然科学版）》;第42卷(第5期);全文 *
"基于深度学习的电气设备图像实例分割算法研究";李雷垚;《中国优秀硕士学位论文全文数据库 工程科技Ⅱ辑》;全文 *
"大型水轮发电机转子金属性一点接地真机故障模拟与快速查找试验";郭钰静;《水电与新能源》;第33卷(第2期);全文 *

Also Published As

Publication number Publication date
CN115393195A (en) 2022-11-25

Similar Documents

Publication Publication Date Title
CN111243032B (en) Full-automatic detection method for checkerboard corner points
CN111260788B (en) Power distribution cabinet switch state identification method based on binocular vision
CN112837257B (en) Curved surface label splicing detection method based on machine vision
CN111307818A (en) Visual online detection method for laser welding spot of lithium battery tab
CN110033407B (en) Shield tunnel surface image calibration method, splicing method and splicing system
CN111275759A (en) Transformer substation disconnecting link temperature detection method based on unmanned aerial vehicle double-light image fusion
CN107239748A (en) Robot target identification and localization method based on gridiron pattern calibration technique
CN111243033A (en) Method for optimizing external parameters of binocular camera
CN111508032B (en) Method for sorting feature points in camera calibration process
CN110570422B (en) Capsule defect visual detection method based on matrix analysis
CN103902953B (en) A kind of screen detecting system and method
CN110838086B (en) Outdoor image splicing method based on correlation template matching
CN111461963A (en) Fisheye image splicing method and device
CN111243034A (en) Panoramic auxiliary parking calibration method, device, equipment and storage medium
CN108709500B (en) Circuit board element positioning and matching method
CN111915485A (en) Rapid splicing method and system for feature point sparse workpiece images
CN114331835A (en) Panoramic image splicing method and device based on optimal mapping matrix
CN115393195B (en) Generator stator core infrared image panorama stitching method based on visible light assistance
CN111131801A (en) Projector correction system and method and projector
CN111325218B (en) Hog feature detection and matching method based on light field image
CN117095417A (en) Screen shot form image text recognition method, device, equipment and storage medium
CN104408430B (en) License plate positioning method and device
CN113554672B (en) Camera pose detection method and system in air tightness detection based on machine vision
CN112215319B (en) Two-dimensional code of color mark characteristic graph and identification method thereof
CN112184533B (en) Watermark synchronization method based on SIFT feature point matching

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