CN115482206A - Method for detecting overall size of ultra-large type reinforcing steel bar net piece - Google Patents
Method for detecting overall size of ultra-large type reinforcing steel bar net piece Download PDFInfo
- Publication number
- CN115482206A CN115482206A CN202211061272.3A CN202211061272A CN115482206A CN 115482206 A CN115482206 A CN 115482206A CN 202211061272 A CN202211061272 A CN 202211061272A CN 115482206 A CN115482206 A CN 115482206A
- Authority
- CN
- China
- Prior art keywords
- image
- mesh
- steel bar
- pixel points
- reinforcing
- 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
- 229910001294 Reinforcing steel Inorganic materials 0.000 title claims abstract description 39
- 238000000034 method Methods 0.000 title claims abstract description 31
- 230000003014 reinforcing effect Effects 0.000 claims description 70
- 229910000831 Steel Inorganic materials 0.000 claims description 36
- 239000010959 steel Substances 0.000 claims description 36
- 238000001914 filtration Methods 0.000 claims description 12
- 238000007781 pre-processing Methods 0.000 claims description 6
- 238000012935 Averaging Methods 0.000 claims description 5
- 239000011159 matrix material Substances 0.000 claims description 5
- 230000003287 optical effect Effects 0.000 claims description 4
- 239000003086 colorant Substances 0.000 claims description 3
- 230000000007 visual effect Effects 0.000 claims description 2
- 230000002787 reinforcement Effects 0.000 claims 1
- 238000001514 detection method Methods 0.000 abstract description 17
- 230000008030 elimination Effects 0.000 abstract description 3
- 238000003379 elimination reaction Methods 0.000 abstract description 3
- 238000012545 processing Methods 0.000 description 9
- 238000006243 chemical reaction Methods 0.000 description 7
- 238000004519 manufacturing process Methods 0.000 description 5
- 238000003466 welding Methods 0.000 description 2
- 238000005452 bending Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000012217 deletion Methods 0.000 description 1
- 230000037430 deletion Effects 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000010008 shearing Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0004—Industrial image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/20—Image enhancement or restoration using local operators
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/70—Denoising; Smoothing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/187—Segmentation; Edge detection involving region growing; involving region merging; involving connected component labelling
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/60—Analysis of geometric attributes
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/90—Determination of colour characteristics
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Quality & Reliability (AREA)
- Geometry (AREA)
- Image Analysis (AREA)
Abstract
The invention relates to a method for detecting the overall size of an ultra-large reinforcing steel bar mesh, which comprises the steps of flatly placing the ultra-large reinforcing steel bar mesh on a bearing platform with a white background, carrying out image acquisition on an industrial camera carried on a portal frame in a mode of being vertical to the bearing platform and looking down, and arranging an image acquisition environment according to a working distance; obtaining the relation between the real size and the number of pixel points; through the steps of image graying, denoising pretreatment, reinforcing steel bar mesh area growth, reinforcing steel bar transverse rib elimination, calculation of the whole size of the reinforcing steel bar mesh and the like, the whole size detection of the reinforcing steel bar mesh is creatively realized, and the speed and the precision meet the actual generation requirement.
Description
Technical Field
The invention belongs to the technical field of engineering measurement, and particularly relates to an image recognition-based method for detecting the overall size of an ultra-large reinforcing mesh.
Background
A simply supported box girder in railway engineering generally adopts a steel bar net piece as a skeleton support, the manufacturing process of the steel bar net piece mainly depends on manual processing, sorting, placing and binding, and the whole process consumes time and labor and is seriously dependent on manpower. Particularly, when the steel bar framework main body is composed of ultra-large steel bar meshes, the maximum self weight of each mesh is up to 10 tons, and the processing by only manpower is particularly difficult. Therefore, special reinforcing mesh welding equipment is adopted for production. The whole size of the reinforcing mesh can deviate from the design requirement to a certain extent due to the straightening, shearing, welding, bending and other links in the production process of the reinforcing mesh, so that the normal use and the function of the simply supported box girder are influenced, and the great potential safety hazard is caused.
The upper limit of the overall size of the ultra-large reinforcing mesh can reach 32m multiplied by 10m, if the manual detection is relied on, the time and the labor are wasted, and the detection precision is insufficient. The traditional small object detection scheme is difficult to apply, the size detection research of the ultra-large object is less, because in the actual production of a factory, the field light source is seriously insufficient, the arrangement mode of image acquisition needs to be strictly calculated, the image quality and the identification precision are directly influenced, and the 3D point cloud modeling method has too long detection period and insufficient precision; therefore, the invention is particularly important for quickly and accurately detecting the overall size of the ultra-large reinforcing mesh.
Disclosure of Invention
Aiming at the defects in the prior art, the invention aims to provide a quick and accurate method for detecting the overall size of the ultra-large reinforcing mesh.
The technical scheme adopted by the invention for solving the technical problems is as follows:
a method for detecting the whole size of an ultra-large reinforcing mesh comprises the following steps:
step one, image acquisition
Flatly placing the ultra-large reinforcing steel bar mesh on a bearing platform with a white background, carrying out image acquisition on an industrial camera loaded on a portal frame in a mode of being vertical to the bearing platform and looking down, and recording the vertical distance from the industrial camera to the ultra-large reinforcing steel bar mesh as a working distance, wherein D represents the vertical distance; the working distance D = the long side of the visual field multiplied by the focal length/chip length, and the image acquisition environment is arranged according to the working distance; obtaining the relation between the real size and the number of pixel points, namely the real size/optical magnification = the number of corresponding pixel points;
step two, graying the image
The colors of all pixel points of the reinforcing steel bar mesh image are represented by an RGB color system, wherein R, G and B respectively represent the brightness values of a Red color channel, a Green color channel and a Blue color channel; carrying out image graying by utilizing the weighted average value to complete image graying processing, and setting the steel bar mesh image after the image graying processing as a steel bar mesh gray image, wherein the brightness value of the steel bar mesh image is G, G = R multiplied by 0.30+ G multiplied by 0.59+ B multiplied by 0.11;
step three, denoising pretreatment
Carrying out Gaussian filtering denoising treatment by using a 9 multiplied by 9 Gaussian kernel template, eliminating or reducing noise generated in a complex shooting environment of a factory, and setting an image after noise elimination as a steel mesh pretreatment image;
step four, growing the reinforcing mesh area
Calculating the gradient amplitude and the azimuth angle of all pixel points in the pre-processed image of the reinforcing mesh, obtaining the position coordinate, the scale (namely the gradient amplitude) and the direction (namely the azimuth angle) of each pixel point, and traversing all the images of the pre-processed image of the reinforcing meshPrime dots, sequentially recorded as A 1 、A 2 、…A N Wherein N represents the total number of all pixel points, all pixel points are sorted from small to large according to the gradient amplitude, and all the sorted pixel points are renamed to be k 1 、k 2 、…k i 、…、k N Selecting the pixel point with the maximum gradient amplitude as the starting point of region growth, namely using k N Selecting 3 x 3 grids as starting points of region growth by taking the representative pixel points as centers to carry out region growth, and enabling k to be N The pixel points represented and the gradient descent in the 3 x 3 grid are most obvious and the azimuth angle and k max Connecting the pixels with the azimuth angle difference within +/-22.5 degrees, selecting a 3 multiplied by 3 grid by taking the pixel as a new center to carry out region growth, and repeating the process until no pixel with the azimuth angle difference within +/-22.5 degrees exists in the 3 multiplied by 3 grid;
then, from k N-1 Starting with k N-1 Selecting a 3 multiplied by 3 grid for region growth by taking the represented pixel points as centers, and repeating the process of region growth until all the pixel points finish the region growth so as to form a line segment;
considering that the reinforcing mesh consists of transverse ribs and longitudinal ribs, only horizontal line segments and vertical line segments need to be reserved, the slope of all line segments with the growing completion of the region is calculated, the line segments with the slope of 0 and infinity are reserved, otherwise, after irrelevant line segments are eliminated, the number of all remaining line segments is set as M, and the number of pixel points contained in each line segment is sequentially recorded as T 1 、T 2 、…、T i 、…T M ；
Step five, removing the transverse ribs of the reinforcing steel bars
Setting the actual size value range of the reinforcing steel bar transverse rib, obtaining the pixel number range of the transverse rib according to the relation between the actual size and the pixel number, setting the upper limit of the pixel number range of the transverse rib as S, and setting the pixel number T contained in each line segment in the preprocessed image of the reinforcing steel bar mesh obtained in the fourth step i If T is compared with the upper limit S i If the image is less than S, deleting the steel bar transverse rib by the line segment, and setting the image without the steel bar transverse rib as the final image of the steel bar net sheet;
step six, calculating the overall size of the reinforcing mesh
Through the fifth step, the final image of the reinforcing mesh only comprises transverse ribs and vertical ribs of the reinforcing mesh, the reinforcing steel bars in the length direction of the ultra-large reinforcing mesh are set as longitudinal ribs, the reinforcing steel bars in the width direction are transverse ribs, the size of the longitudinal ribs of the ultra-large reinforcing mesh is much larger than that of the transverse ribs, the number of corresponding pixel points contained in all remaining line segments after the transverse ribs of the reinforcing steel bars are removed is obtained, the remaining line segments are divided into two types, namely the transverse ribs and the longitudinal ribs according to two trends of the number of the pixel points, one type with obviously large number of the pixel points is classified as the longitudinal ribs, and the other type with obviously small number of the pixel points is classified as the transverse ribs;
the method comprises the steps of averaging the number of pixels contained in each line segment in a longitudinal rib of the reinforcing mesh by utilizing the relation between the true size and the number of the pixels, converting the average number of the pixels into the actual length of the reinforcing mesh, averaging the number of the pixels contained in each line segment in a transverse rib of the reinforcing mesh, converting the average number of the pixels into the actual width of the reinforcing mesh, and obtaining the overall size of the ultra-large reinforcing mesh.
The upper limit of the overall size of the ultra-large reinforcing steel mesh is 32m multiplied by 10m, and the actual size range of the transverse ribs of the reinforcing steel bars is 6-8mm.
Compared with the prior art, the invention has the beneficial effects that:
1. aiming at the problem that the actual size cannot be directly obtained from the steel bar mesh image, the invention can accurately calculate the actual size from the steel bar mesh image by setting an experimental environment and a size conversion calculation formula, thereby realizing size conversion and ensuring detection precision;
2. aiming at the problems that the environment of a reinforcing steel bar processing site is complex, and the detection precision is influenced by excessive image noise, the invention reduces the possibility of noise generation to the greatest extent from the source by arranging an experimental environment and a white background bottom plate, and can reduce the probability of false detection;
3. aiming at the problems that the Gaussian filtering preprocessing time of the reinforcing steel bar mesh image is too long, and the detection speed cannot meet the actual requirement, the processing time of an average reinforcing steel bar mesh gray level image is reduced to 1s by improving the Gaussian filtering kernel template, so that the operation speed is fastest on the premise of removing noise as much as possible;
4. aiming at the problems that the diameter size of the steel bar of the ultra-large steel bar mesh is large and the identification precision is influenced by the transverse ribs of the steel bar, the transverse ribs of the steel bar are removed according to the conversion relation between the pixel points and the actual size, so that the detection precision is improved;
5. aiming at the problem that the size detection method of the ultra-large reinforcing mesh sheet is less, the invention creatively realizes the whole size detection of the reinforcing mesh sheet through the steps of reinforcing mesh sheet region growth, reinforcing steel bar transverse rib elimination, size calculation and the like, and the speed and the precision both meet the actual generation requirement.
Drawings
FIG. 1 is an overall flow diagram of the process of the present invention;
FIG. 2 is a flow chart of the conversion between the reinforcing mesh image and the overall size according to the method of the present invention;
FIG. 3 is a reinforcing mesh image acquisition environment of the method of the present invention;
FIG. 4 is an image of a rebar mesh of an embodiment of the method of the present invention;
FIG. 5 is a gray scale image of a rebar mesh in accordance with an embodiment of the method of the present invention;
FIG. 6 is a pre-processed image of a rebar mesh in accordance with an embodiment of the method of the present invention;
FIG. 7 is a final image of a rebar mesh of an embodiment of the method of the present invention;
Detailed Description
The present invention is further explained with reference to the following examples and drawings, but the scope of the present invention is not limited thereto.
The invention relates to a quick and accurate method for detecting the overall size of an ultra-large reinforcing mesh, which comprises the following steps:
step one, image acquisition
The image acquisition equipment comprises portal frame and the cushion cap that has white background, and wherein, the portal frame is used for carrying on the industry camera, and the cushion cap is used for placing super large-scale reinforcing bar net piece. When the image acquisition work is carried out, the ultra-large steel bar mesh is flatly placed on a bearing platform with a white background, an industrial camera carried on a portal frame is used for carrying out image acquisition in a mode of being vertical to the bearing platform and looking down, the vertical distance from the industrial camera to the ultra-large steel bar mesh is recorded as a working distance and is represented by D, the industrial camera parameters are consulted, the resolution ratio is recorded as a multiplied by b, the pixel size is recorded as c multiplied by D, and the focal length is recorded as f, and the table 1 shows that the image acquisition work is carried out.
TABLE 1 Industrial Camera parameters
| Serial number | Industrial camera parameters | Specific numerical value |
| 1 | Resolution ratio | a×b |
| 2 | Size of picture element | c×d |
| 3 | Focal length | f |
Step two, calculating the size conversion relation
The specific numerical value of the working distance D is calculated according to the following calculation formula.
Chip size = resolution × pixel size
Optical magnification = long side of field of view/length of camera chip = working distance/focal length
Working distance = long side of field of view × focal length/chip length
True size/optical magnification = number of corresponding pixel points
And (3) adjusting the exposure and the gain of the industrial camera so as to shoot the high-brightness steel mesh image, and arranging the image acquisition environment in the step (1) according to the working distance.
Step three, graying the image
The colors of all pixel points of the reinforcing mesh image are represented by an RGB color system, wherein R, G and B respectively represent the brightness values of a Red (Red) color channel, a Green (Green) color channel and a Blue (Blue) color channel. Considering that the color of the reinforcing steel bar mesh is substantially gray, the color is single and the change is not obvious, image graying is performed by using a weighted average value to complete image preprocessing, and the reinforcing steel bar mesh image after image preprocessing is a reinforcing steel bar mesh gray image with a brightness value of G, G = R × 0.30+ G × 0.59+ B × 0.11.
Step four, improving Gaussian filtering to remove noise
In order to eliminate or reduce noise generated in a complex shooting environment of a factory, gaussian filtering is used for linear smooth filtering, and an image with the noise removed is set as a steel mesh preprocessing image. The traditional Gaussian filtering method adopts a 3 x 3 Gaussian kernel template, and through experimental tests, the processing time of an average reinforcing mesh gray image is about 5s, but aiming at the characteristics of overlarge size and overhigh shooting precision of a reinforcing mesh, the Gaussian kernel template cannot meet the rapid detection requirement of actual production, and the Gaussian filtering method is improved to generate a 9 x 9 Gaussian kernel template, as shown in table 2.
TABLE 2 Gauss filtering 9X 9 Gauss kernel template
| (x-4,y+4) | (x-3,y+4) | (x-2,y+4) | (x-1,y+4) | (x,y+4) | (x+1,y+4) | (x+2,y+4) | (x+3,y+4) | (x+4,y+4) |
| (x-4,y+3) | (x-3,y+3) | (x-2,y+3) | (x-1,y+3) | (x,y+3) | (x+1,y+3) | (x+2,y+3) | (x+3,y+3) | (x+4,y+3) |
| (x-4,y+2) | (x-3,y+2) | (x-2,y+2) | (x-1,y+2) | (x,y+2) | (x+1,y+2) | (x+2,y+2) | (x+3,y+2) | (x+4,y+2) |
| (x-4,y+1) | (x-3,y+1) | (x-2,y+1) | (x-1,y+1) | (x,y+1) | (x+1,y+1) | (x+2,y+1) | (x+3,y+1) | (x+4,y+1) |
| (x-4,y) | (x-3,y) | (x-2,y) | (x-1,y) | (x,y) | (x+1,y) | (x+2,y) | (x+3,y) | (x+4,y) |
| (x-4,y-1) | (x-3,y-1) | (x-2,y-1) | (x-1,y-1) | (x,y-1) | (x+1,y-1) | (x+2,y-1) | (x+3,y-1) | (x+4,y-1) |
| (x-4,y-2) | (x-3,y-2) | (x-2,y-2) | (x-1,y-2) | (x,y-2) | (x+1,y-2) | (x+2,y-2) | (x+3,y-2) | (x+4,y-2) |
| (x-4,y-3) | (x-3,y-3) | (x-2,y-3) | (x-1,y-3) | (x,y-3) | (x+1,y-3) | (x+2,y-3) | (x+3,y-3) | (x+4,y-3) |
| (x-4,y-4) | (x-3,y-4) | (x-2,y-4) | (x-1,y-4) | (x,y-4) | (x+1,y-4) | (x+2,y-4) | (x+3,y-4) | (x+4,y-4) |
Wherein, (x, y) is the pixel point coordinate of reinforcing bar net piece grey scale image, and its gaussian value that corresponds is h (x, y), adopts the gaussian formula to advance gaussian filtering:
considering that the standard deviation sigma influences the operation speed of the gaussian filter, in order to further increase the detection speed, the standard deviation sigma is adjusted, the optimum standard deviation sigma suitable for the gaussian kernel template with the standard deviation sigma of 9 × 9 is 1.5, the processing time of the average reinforcing mesh gray level image is reduced to 1s, the operation speed is fastest on the premise that noise is removed as far as possible, and at the moment, the weight matrix of the gaussian filter is shown in table 3.
TABLE 3 weight matrix corresponding to Gauss filter 9X 9 Gauss kernel template
Fifthly, growing the reinforcing mesh area
Calculating the gradient amplitude and the azimuth angle of all pixel points in the preprocessed image of the reinforcing mesh, wherein the formula is as follows:
the image gradient calculation formula in the horizontal direction is as follows:
the y-axis image gradient calculation formula of the matrix a is as follows:
gradient amplitude of the image pixel points preprocessed by the reinforcing mesh is as follows:
azimuth of image pixel point of reinforcing bar net piece preliminary treatment:
wherein, each pixel point can obtain position coordinates (i.e., (x, y)), dimensions (i.e., gradient amplitude) and direction (i.e., azimuth angle), and all pixel points of the preprocessed image of the steel bar mesh are traversed and recorded as A in sequence 1 、A 2 、…A N Wherein N represents the total number of all pixel points, all pixel points are sorted from small to large according to gradient amplitudes, and all the sorted pixel points are renamed to be k 1 、k 2 、…k i 、…、k N Selecting the pixel point with the maximum gradient amplitude as the starting point of region growth, namely using k N Selecting 3 x 3 grids as starting points of region growth by taking the representative pixel points as centers to carry out region growth, and enabling k to be N The pixel points represented are the most significant with gradient descent in the 3 x 3 grid and the azimuth and k max Connecting the pixels with the azimuth angle difference within +/-22.5 degrees, selecting a 3 multiplied by 3 grid by taking the pixel as a new center to carry out region growth, and repeating the process until no pixel with the azimuth angle difference within +/-22.5 degrees exists in the 3 multiplied by 3 grid;
then, from k N-1 Starting with k N-1 Selecting a 3 multiplied by 3 grid for region growth by taking the represented pixel points as centers, and repeating the process of region growth until all the pixel points finish the region growth so as to form a line segment;
considering that the reinforcing steel bar mesh consists of transverse bars and longitudinal bars, only horizontal line segments and vertical line segments need to be reserved, the reinforcing steel bars in the length direction of the ultra-large reinforcing steel bar mesh are assumed to be longitudinal bars, the reinforcing steel bars in the width direction are transverse bars, all line segment slopes of area growth completion in the fifth step are calculated, line segments with the slopes of 0 and infinity are reserved, irrelevant line segments are removed, the number of all remaining line segments at the moment is set to be M, and the number of pixel points contained in each line segment is sequentially recorded as T 1 、T 2 、…、T i 、…T M 。
Sixth, removing the transverse ribs of the reinforcing steel bars
Considering that the diameter of the steel bar of the ultra-large steel bar mesh is generally 12-16mm, the size conversion relation in the second step can know the pixel number range (the range of the transverse rib is generally 6-8 mm) of the corresponding transverse rib, the upper limit of the pixel number range of the transverse rib is S, and the upper limit of the pixel number range of the transverse rib is S, so that the accuracy of the overall size detection of the steel bar mesh can be directly influenced i Deletion of < S line segment (i.e. steel)Rib cross rib), and setting the image of the steel bar cross rib to be eliminated as the final image of the steel bar mesh.
Step seven, calculating the overall size of the reinforcing mesh
After the processing of the sixth step, the final image of the reinforcing mesh only contains the transverse bars and the longitudinal bars of the reinforcing mesh, and due to the characteristic that the longitudinal bars of the ultra-large reinforcing mesh are much larger than the transverse bars, the residual line segments determined after the transverse ribs of the reinforcing mesh are removed in the sixth step are divided into two categories, namely the transverse bars and the longitudinal bars, the pixel number determines the length of the reinforcing mesh, the lengths of all the transverse bars are similar, the lengths of all the longitudinal bars are similar, the category with longer length (i.e. the category with more pixel points) is classified as the longitudinal bars, and the category with short length (i.e. the category with less pixel points) is classified as the transverse bars.
And (3) averaging the number of pixel points contained in each segment in the longitudinal bars of the reinforcing mesh by using the size conversion relation calculation formula in the second step so as to convert the average number of pixel points into the actual length of the reinforcing mesh.
In this embodiment, concrete values of the camera parameters used for the image acquisition of the steel mesh sheet are shown in table 4, and the size detection result is shown in table 5.
TABLE 4 reinforcing mesh image acquisition uses camera parameters
TABLE 5 output results
Nothing in this specification is said to apply to the prior art.
Claims (4)
1. A method for detecting the whole size of an ultra-large reinforcing mesh comprises the following steps:
step one, image acquisition
Flatly placing the ultra-large reinforcing steel bar mesh on a bearing platform with a white background, carrying out image acquisition on an industrial camera loaded on a portal frame in a mode of being vertical to the bearing platform and looking down, and recording the vertical distance from the industrial camera to the ultra-large reinforcing steel bar mesh as a working distance, wherein D represents the vertical distance; the working distance D = the long side of the visual field multiplied by the focal length/chip length, and the image acquisition environment is arranged according to the working distance; obtaining the relation between the real size and the number of pixel points, namely the real size/optical magnification = the number of corresponding pixel points;
step two, graying the image
The colors of all pixel points of the reinforcing steel bar mesh image are represented by an RGB color system, wherein R, G and B respectively represent the brightness values of a Red color channel, a Green color channel and a Blue color channel; carrying out image graying by utilizing the weighted average value to complete image graying, and setting the steel bar mesh image subjected to image graying as a steel bar mesh gray image, wherein the brightness value of the steel bar mesh gray image is G, and G = R × 0.30+ G × 0.59+ B × 0.11;
step three, denoising pretreatment
Carrying out Gaussian filtering denoising treatment by using a 9 multiplied by 9 Gaussian kernel template, eliminating or reducing noise generated in a complex shooting environment of a factory, and setting an image after the noise is removed as a steel bar mesh preprocessing image;
step four, growing the reinforcing mesh area
Calculating the gradient amplitude and the azimuth angle of all pixel points in the pre-processed image of the reinforcing mesh, wherein the position coordinate, the scale (namely the gradient amplitude) and the direction (namely the azimuth angle) of each pixel point can be obtained, traversing all the pixel points of the pre-processed image of the reinforcing mesh, and sequentially recording the position coordinate, the scale (namely the gradient amplitude) and the direction (namely the azimuth angle) as A 1 、A 2 、…A N Wherein N represents the total number of all pixel points, all pixel points are sorted from small to large according to gradient amplitudes, and all the sorted pixel points are renamed to be k 1 、k 2 、…k i 、…、k N Selecting the pixel point with the maximum gradient amplitude as the starting point of region growth, namely using k N Selecting 3 x 3 grids as starting points of region growth by taking the representative pixel points as centers to carry out region growth, and enabling k to be N The pixel points represented and the gradient descent in the 3 x 3 grid are most obvious and the azimuth angle and k max Connecting the pixel points with the azimuth difference within +/-22.5 degrees, selecting a 3 x 3 grid as a new center for region growth by using the pixel points, and repeating the process until no pixel points with the azimuth difference within +/-22.5 degrees exist in the 3 x 3 grid;
then, from k N-1 At the beginning, with k N-1 Selecting a 3 multiplied by 3 grid for region growth by taking the represented pixel points as centers, and repeating the process of region growth until all the pixel points finish the region growth so as to form a line segment;
considering that the reinforcing mesh consists of transverse ribs and longitudinal ribs, only horizontal line segments and vertical line segments need to be reserved, the slopes of all the line segments with the growing regions finished are calculated, the line segments with the slopes of 0 and infinity are reserved, otherwise, after irrelevant line segments are eliminated, the number of all the remaining line segments is set to be M, and the number of pixel points contained in each line segment is sequentially recorded as T 1 、T 2 、…、T i 、…T M ；
Step five, removing the transverse ribs of the reinforcing steel bars
Setting the actual size value range of the reinforcing steel bar transverse rib, obtaining the pixel number range of the transverse rib according to the relation between the actual size and the pixel number, setting the upper limit of the pixel number range of the transverse rib as S, and setting the pixel number T contained in each line segment in the preprocessed image of the reinforcing steel bar mesh obtained in the fourth step i If T is compared with the upper limit S i If the image is less than S, deleting the steel bar transverse rib by the line segment, and setting the image without the steel bar transverse rib as the final image of the steel bar net sheet;
step six, calculating the overall size of the reinforcing mesh
Through the fifth step, the final image of the reinforcing mesh only comprises transverse ribs and vertical ribs of the reinforcing mesh, the reinforcing steel bars in the length direction of the ultra-large reinforcing mesh are set as longitudinal ribs, the reinforcing steel bars in the width direction are transverse ribs, the size of the longitudinal ribs of the ultra-large reinforcing mesh is much larger than that of the transverse ribs, the number of corresponding pixel points contained in all remaining line segments after the transverse ribs of the reinforcing steel bars are removed is obtained, the remaining line segments are divided into two types, namely the transverse ribs and the longitudinal ribs according to two trends of the number of the pixel points, one type with obviously large number of the pixel points is classified as the longitudinal ribs, and the other type with obviously small number of the pixel points is classified as the transverse ribs;
the method comprises the steps of averaging the number of pixels contained in each line segment in a longitudinal rib of the reinforcing mesh by utilizing the relation between the true size and the number of the pixels, converting the average number of the pixels into the actual length of the reinforcing mesh, averaging the number of the pixels contained in each line segment in a transverse rib of the reinforcing mesh, converting the average number of the pixels into the actual width of the reinforcing mesh, and obtaining the overall size of the ultra-large reinforcing mesh.
2. The method for detecting the overall size of the ultra-large reinforcing mesh piece according to claim 1, wherein the upper limit of the overall size of the ultra-large reinforcing mesh piece is 32m x 10m, and the actual size range of the transverse reinforcing rib is 6-8mm.
3. The method for detecting the overall size of the ultra-large reinforcing mesh piece according to claim 1, wherein the 9 x 9 Gaussian kernel template is as follows:
wherein, (x, y) is the pixel point coordinate of reinforcing bar net piece grey scale image, and its gaussian value that corresponds is h (x, y), adopts the gaussian formula to advance gaussian filtering:
the standard deviation σ is 1.5, and the gaussian filter weight matrix is:
4. the method for detecting the overall size of the ultra-large reinforcing mesh piece according to claim 3, wherein the calculation process of the gradient amplitude and the azimuth angle of the pixel point is as follows:
image gradient g in horizontal direction x (x, y) the formula is as follows:
image gradient g in y-axis direction of matrix A y (x, y) the formula is as follows:
gradient amplitude g (x, y) of reinforcing bar net piece preprocessing image pixel point:
azimuth angle theta (x, y) of image pixel points of reinforcement mesh pretreatment:
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211061272.3A CN115482206A (en) | 2022-09-01 | 2022-09-01 | Method for detecting overall size of ultra-large type reinforcing steel bar net piece |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211061272.3A CN115482206A (en) | 2022-09-01 | 2022-09-01 | Method for detecting overall size of ultra-large type reinforcing steel bar net piece |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115482206A true CN115482206A (en) | 2022-12-16 |
Family
ID=84421773
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211061272.3A Pending CN115482206A (en) | 2022-09-01 | 2022-09-01 | Method for detecting overall size of ultra-large type reinforcing steel bar net piece |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115482206A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116524004A (en) * | 2023-07-03 | 2023-08-01 | 中国铁路设计集团有限公司 | Method and system for detecting size of steel bar based on HoughLines algorithm |
-
2022
- 2022-09-01 CN CN202211061272.3A patent/CN115482206A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116524004A (en) * | 2023-07-03 | 2023-08-01 | 中国铁路设计集团有限公司 | Method and system for detecting size of steel bar based on HoughLines algorithm |
| CN116524004B (en) * | 2023-07-03 | 2023-09-08 | 中国铁路设计集团有限公司 | Method and system for detecting size of steel bar based on HoughLines algorithm |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112651968B (en) | Wood board deformation and pit detection method based on depth information | |
| CN110751604B (en) | Online detection method for weld defects of steel pipe based on machine vision | |
| CN108596880A (en) | Weld defect feature extraction based on image procossing and welding quality analysis method | |
| CN113344956B (en) | Ground feature contour extraction and classification method based on unmanned aerial vehicle aerial photography three-dimensional modeling | |
| CN112613097A (en) | BIM rapid modeling method based on computer vision | |
| CN107610092B (en) | Pavement crack dynamic detection method based on video stream | |
| CN113222992B (en) | Crack characteristic characterization method and system based on multi-fractal spectrum | |
| CN112258446A (en) | Industrial part defect detection method based on improved YOLO algorithm | |
| CN115482206A (en) | Method for detecting overall size of ultra-large type reinforcing steel bar net piece | |
| CN111104860A (en) | Unmanned aerial vehicle water quality chromaticity monitoring method based on machine vision | |
| CN114219773B (en) | Pre-screening and calibrating method for bridge crack detection data set | |
| CN104008528A (en) | Inhomogeneous light field underwater target detection image enhancing method based on threshold segmentation | |
| CN113252103A (en) | Method for calculating volume and mass of material pile based on MATLAB image recognition technology | |
| CN113327255A (en) | Power transmission line inspection image processing method based on YOLOv3 detection, positioning and cutting and fine-tune | |
| CN112465809A (en) | Mold defect detection method based on image recognition, computer-readable storage medium and device | |
| CN111401358A (en) | Instrument dial plate correction method based on neural network | |
| CN108665468B (en) | Device and method for extracting tangent tower insulator string | |
| CN115602093A (en) | Method, system and equipment for performing Demura compensation based on white picture | |
| CN117274240A (en) | Bearing platform foundation concrete surface crack identification method | |
| CN111724354A (en) | Image processing-based method for measuring spike length and small spike number of multiple wheat | |
| CN105550750B (en) | The method for improving convolutional neural networks accuracy of identification | |
| CN111429437B (en) | Image non-reference definition quality detection method for target detection | |
| CN116109633B (en) | Window detection method and device for bearing retainer | |
| CN110717592B (en) | Method for calculating space distribution of irradiance of surface of photovoltaic power station | |
| CN116542910A (en) | Stamping part round hole positioning and deformation detection method based on sub-pixel edge extraction |
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 |