CN114331973A - Steel structure information extraction method suitable for oil-gas module manufacturing process - Google Patents

Abstract

The invention discloses a steel structure information extraction method suitable for an oil-gas module manufacturing process, which comprises the following steps: the method comprises the steps of laser engraving a steel structure production serial number and storing the steel structure production serial number in a database, shooting the steel structure surface production serial number by a CCD vision camera, introducing a filtering algorithm into python to perform noise reduction processing on a picture, performing image alignment processing in python, performing training of an image production serial number prediction model by using an MLP neural network algorithm, recognizing on-site steel structure production serial number information, sending the recognition information to a server through a network, and displaying the information on a display screen. Compared with the traditional method, the method can effectively solve the problems of complex information searching work, less information acquisition amount, poor information integrity and the like in the oil-gas module manufacturing process, and is efficient, reliable, strong in universality and strong in adaptability.

Description

Steel structure information extraction method suitable for oil-gas module manufacturing process

Technical Field

The invention relates to an information extraction method, in particular to a steel structure information extraction method suitable for an oil-gas module manufacturing process.

Background

The oil gas module is formed by welding various steel structures, the steel structures are various and numerous, and the accurate acquisition and the correspondence of the information of each steel structure are very difficult. In many manufacturing fields, means such as bar codes and RFID labels are adopted to facilitate information query, but in the process of manufacturing oil gas modules, common labels cannot be applied to field operation environments due to extreme climates such as strong wind current, high welding temperature and the like. In the manufacturing process of present oil gas module, on-the-spot engineer still adopts the mode of paper drawing to seek the information of steel construction, seeks inefficiency, and it is limited to seek the information, can not satisfy actual engineering demand completely, needs a high-efficient stable, convenient easy information extraction method in oil gas module manufacturing process urgently.

Disclosure of Invention

The invention aims to overcome the defects of the prior art and provide the steel structure information extraction method which can improve the speed of extracting the basic information of the steel structure, improve the accuracy of information of various parts in field operation, ensure that the oil-gas module manufacturing process is smoothly carried out, adapt to complex operating environments, have strong universality and simple search mode and is suitable for the oil-gas module manufacturing process.

The invention relates to a steel structure information extraction method suitable for an oil-gas module manufacturing process, which comprises the following steps of:

firstly, when various steel structures of an oil-gas module are built, a unique production serial number is engraved on the surface of the steel structure by utilizing laser, and the production serial number and various basic information of the steel structure are stored in a database of a server end in a key-value pair mode, wherein the production serial number is a key, and the various basic information is a value;

step two, photographing the production serial number of the surface of the steel structure by using a CCD (charge coupled device) vision camera, enabling a lens plane of the CCD vision camera to be parallel to the surface of the steel structure during photographing, and transmitting a photographed picture to a computer through an industrial Ethernet;

opening a picture shot by a CCD visual camera in a python programming environment, and sequentially introducing a median filtering algorithm and a wiener filtering algorithm to perform noise reduction on the picture to obtain a noise reduction image;

fourthly, performing alignment processing on the noise reduction processing image in a computer, wherein the specific process is as follows:

step one, a numpy module is led in the Python, discrete point coordinate extraction operation is carried out on the noise reduction processing image obtained in the step two, and a point A with the maximum horizontal coordinate in all discrete points is obtained₁The point A having the smallest abscissa₂Point B with the largest ordinate₁Point B having the smallest ordinate₂Creating a first bounding box and a second bounding box according to four points, wherein the two bounding boxes are rectangular and establishing a linear equation of four edges of each bounding box; wherein the first and second sides of the first bounding box are parallel to the straight line A₁A₂And the first edge passes through point B₁The second side passes through the point B₂The third edge passes through point A₁The fourth side passes through point A₂(ii) a The first and second sides of the second bounding box are parallel to the straight line B₁B₂And the first edge passes through point A₁The second side passing through point A₂The third edge passes through point B₁The fourth side passing through B₂

Step two, introducing a K-means module into Python, respectively introducing the four-edge equation of the first bounding box and the four-edge equation of the second bounding box obtained in the step one, and calculating by using a clustering algorithm to obtain the number n of discrete points in the first bounding box₁Number of discrete points n in the second enclosure₂Comparison of n₁And n₂Selecting the bounding box containing the largest number of discrete points as a serial number bounding box and solving the slope of the long side of the serial number bounding box;

thirdly, combining the first step and the second step to calculate to obtain the inclination angle theta of the sequence number bounding box and solve the coordinate of the central point of the sequence number bounding box, wherein the theta indicates the included angle between the long edge of the sequence number bounding box and the x axis;

fourthly, clockwise rotating the noise reduction processing image used in the first step around the central point of the sequence number bounding box according to the inclination angle theta to obtain an alignment processing image;

step five, training an image production sequence number prediction model by using an MLP neural network algorithm, wherein the specific process is as follows:

step one, repeating the step two to the step four for a plurality of times to obtain a large number of alignment processing images as image samples, and manually determining the production serial number corresponding to each image sample;

step two, introducing a neural network module into python, extracting an image sample obtained in the step one, and identifying the production sequence number by using an MPL neural network algorithm to obtain an identification result of the image sample;

thirdly, judging whether the identification result is matched with the manually identified production sequence number result, and returning to the second step to identify the production sequence number of the image sample by reusing the MPL neural network algorithm if the identification result is failed to be matched with the manually identified production sequence number result; if the matching is successful, returning to the second step to extract the next image sample for identifying the production serial number until all the image samples can be successfully identified to obtain an image production serial number identification model;

step six, in the actual operation of the oil-gas module, selecting a steel structure, sequentially performing the processes of the step two, the step three and the step four, and calculating the alignment processing image obtained in the step four by using the image production sequence number prediction model obtained in the step five in python to obtain the production sequence number of the steel structure;

and step seven, the computer sends a request to the server through the network, finds the steel structure corresponding to the production sequence in the step six in the database of the server, and displays the basic information of the steel structure on a display screen of the operation site through network transmission, so that the site operation is facilitated.

The invention has the advantages that: the server remote database is adopted to store the basic information of various steel structures, so that the integrity and comprehensiveness of the information are ensured, and the multi-end use in an operation field is also ensured; the production serial number is added on the steel structure in a laser code carving mode, so that the production serial number can resist the environment of a high-temperature manufacturing field and resist corrosion and damage in a service period; the neural network is used to obtain the image character prediction model to complete the identification work of the production serial number, so that the identification efficiency and accuracy can be greatly improved; the main calculation and data processing are concentrated before the operation, the calculated amount is less when the information is searched, and the information searching efficiency is improved; the method can effectively solve the problems of complex information searching work, less information acquisition amount, poor information integrity and the like in the oil-gas module manufacturing process, and is efficient, reliable, strong in universality and strong in adaptability.

Drawings

FIG. 1 is a flow chart of a steel structure information extraction method suitable for use in oil and gas module manufacturing processes;

FIG. 2 is a schematic diagram of data transmission of a steel structure information extraction method suitable for use in oil and gas module manufacturing processes;

FIG. 3 is a schematic diagram of steel structure information extraction method bounding box settings suitable for use in oil and gas module manufacturing processes.

Detailed Description

The invention is described in detail below with reference to the figures and specific embodiments.

The invention relates to a steel structure information extraction method suitable for an oil-gas module manufacturing process, which comprises the following steps:

firstly, when various steel structures of an oil-gas module are built, unique production serial numbers are engraved on the surfaces of the steel structures by utilizing laser, and the production serial numbers and various basic information (such as length information) of the steel structures are stored in a database of a server side in a key-value pair mode, wherein the production serial numbers are keys, and the various basic information are values.

And step two, photographing the production serial number of the surface of the steel structure by using a CCD vision camera, enabling a lens plane of the CCD vision camera to be parallel to the surface of the steel structure during photographing, and transmitting the photographed picture to a computer through an industrial Ethernet.

And step three, opening the picture shot by the CCD visual camera in the python programming environment, and sequentially introducing a median filtering algorithm and a wiener filtering algorithm to perform noise reduction processing on the picture to obtain a noise reduction processing image.

Fourthly, performing alignment processing on the noise reduction processing image in a computer, wherein the specific process is as follows:

step one, a numpy module is led in the Python, discrete point coordinate extraction operation is carried out on the noise reduction processing image obtained in the step two, and a point A with the maximum horizontal coordinate in all discrete points is obtained₁The point A having the smallest abscissa₂Point B with the largest ordinate₁Point B having the smallest ordinate₂And creating a first bounding box and a second bounding box according to four points, wherein the two bounding boxes are rectangular and establishing a linear equation of four sides of each bounding box. Wherein the first and second sides of the first bounding box are parallel to the straight line A₁A₂And the first edge passes through point B₁The second side passes through the point B₂The third edge passes through point A₁The fourth side passes through point A₂. The first and second sides of the second bounding box are parallel to the straight line B₁B₂And the first edge passes through point A₁The second side passing through point A₂The third edge passes through point B₁The fourth side passing through B₂. Therefore, the first, second, third and fourth edge equations of the first bounding box are respectively as follows:

the first, second, third and fourth edge equations of the second bounding box are respectively:

in the formula, x₁、y₁Respectively represent A₁Point abscissa, ordinate, x₂、y₂Respectively represent A₂Point abscissa, ordinate, x₃、y₃Respectively represent B₁Point abscissa, ordinate, x₄、y₄Respectively represent B₂Point abscissa and ordinate; k is a radical of_iARepresenting the slope of the ith edge of the first bounding box. k is a radical of_iBRepresenting the slope of the ith edge of the second bounding box. b_iADenotes the ith edge intercept of the first bounding box, b_iBRepresenting the ith edge intercept of the second bounding box.

Step two, introducing a K-means module into Python, respectively introducing the four-edge equation of the first bounding box and the four-edge equation of the second bounding box obtained in the step one, and calculating by using a clustering algorithm to obtain the number n of discrete points in the first bounding box₁Number of discrete points n in the second enclosure₂Comparison of n₁And n₂Selecting the bounding box containing the largest number of discrete points as the bounding box of the sequence number and solving the slope of the long edge of the bounding box of the sequence number, wherein the process is as follows:

the equation for the ith edge of the numbered bounding box is

y＝k_ix+b_i

Easy to know k_i＝k_iAOr k_iB，b_i＝b_iAOr b_iB。

Setting the intersection point of the first edge and the third edge of the bounding box with the sequence number as C₁₃The intersection point of the first edge and the fourth edge is C₁₄The intersection point of the second side and the third side is C₂₃。C₁₃The coordinates of (A) can be obtained by simultaneous equations of the first and third edges, C₁₄，C₂₃The same process can be used.Comparing line segment C₁₃C₂₃And line segment C₁₃C₁₄The slope of the longer line segment is recorded as k

And thirdly, combining the first step and the second step to calculate to obtain the inclination angle theta of the sequence number bounding box and solve the coordinate of the central point of the sequence number bounding box, wherein the theta indicates the included angle between the long edge of the sequence number bounding box and the x axis.

θ＝arctank

Let the coordinate of the center point of the bounding box with the serial number be (x)₀,y₀) Then there is

In the formula, x₁₄、y₁₄Represents point C₁₄Abscissa, ordinate, x₂₃、y₂₃Represents point C₂₃The abscissa and the ordinate of (a);

fourthly, the central point (x) of the bounding box with the sequence number of the noise reduction processing image used in the first step is surrounded₀,y₀) And rotating clockwise according to the inclination angle theta to obtain an alignment processing image.

Step five, training an image production sequence number prediction model by using an MLP neural network algorithm, wherein the specific process is as follows:

and step one, repeating the step two to the step four for a plurality of times to obtain a large number of alignment processing images as image samples, and manually determining the production serial number corresponding to each image sample.

Step two, introducing a neural network module into python, extracting an image sample obtained in the step one, and identifying the production sequence number by using an MPL neural network algorithm to obtain an identification result of the image sample;

and thirdly, judging whether the identification result is matched with the manually identified production sequence number result, if the matching fails, returning to the second step, and carrying out production sequence number identification on the image sample by using the MPL neural network algorithm again (the model is automatically strengthened in the neural network training process, so that the next identification is closer to correct). And if the matching is successful, returning to the second step to extract the next image sample for identifying the production serial number until all the image samples can be successfully identified to obtain an image production serial number identification model.

And step six, in the actual operation of the oil-gas module, selecting a steel structure, sequentially performing the processes of the step two, the step three and the step four, and calculating the alignment processing image obtained in the step four by using the image production sequence number prediction model obtained in the step five in python to obtain the production sequence number of the steel structure.

And step seven, the computer sends a request to the server through the network, finds the steel structure corresponding to the production sequence in the step six in the database of the server, and displays the basic information of the steel structure on a display screen of the operation site through network transmission, so that the site operation is facilitated.

Claims (1)

1. A steel structure information extraction method suitable for an oil-gas module manufacturing process is characterized by comprising the following steps:

firstly, when various steel structures of an oil-gas module are built, a unique production serial number is engraved on the surface of the steel structure by utilizing laser, and the production serial number and various basic information of the steel structure are stored in a database of a server end in a key-value pair mode, wherein the production serial number is a key, and the various basic information is a value;

step two, photographing the production serial number of the surface of the steel structure by using a CCD (charge coupled device) vision camera, enabling a lens plane of the CCD vision camera to be parallel to the surface of the steel structure during photographing, and transmitting a photographed picture to a computer through an industrial Ethernet;

opening a picture shot by a CCD visual camera in a python programming environment, and sequentially introducing a median filtering algorithm and a wiener filtering algorithm to perform noise reduction on the picture to obtain a noise reduction image;

fourthly, performing alignment processing on the noise reduction processing image in a computer, wherein the specific process is as follows:

first step, in PythonImporting a numpy module, and performing discrete point coordinate extraction operation on the noise reduction processing image obtained in the step two to obtain a point A with the maximum horizontal coordinate in all discrete points₁The point A having the smallest abscissa₂Point B with the largest ordinate₁Point B having the smallest ordinate₂Creating a first bounding box and a second bounding box according to four points, wherein the two bounding boxes are rectangular and establishing a linear equation of four edges of each bounding box; wherein the first and second sides of the first bounding box are parallel to the straight line A₁A₂And the first edge passes through point B₁The second side passes through the point B₂The third edge passes through point A₁The fourth side passes through point A₂(ii) a The first and second sides of the second bounding box are parallel to the straight line B₁B₂And the first edge passes through point A₁The second side passing through point A₂The third edge passes through point B₁The fourth side passing through B₂

Step two, introducing a K-means module into Python, respectively introducing the four-edge equation of the first bounding box and the four-edge equation of the second bounding box obtained in the step one, and calculating by using a clustering algorithm to obtain the number n of discrete points in the first bounding box₁Number of discrete points n in the second enclosure₂Comparison of n₁And n₂Selecting the bounding box containing the largest number of discrete points as a serial number bounding box and solving the slope of the long side of the serial number bounding box;

thirdly, combining the first step and the second step to calculate to obtain the inclination angle theta of the sequence number bounding box and solve the coordinate of the central point of the sequence number bounding box, wherein the theta indicates the included angle between the long edge of the sequence number bounding box and the x axis;

fourthly, clockwise rotating the noise reduction processing image used in the first step around the central point of the sequence number bounding box according to the inclination angle theta to obtain an alignment processing image;

step five, training an image production sequence number prediction model by using an MLP neural network algorithm, wherein the specific process is as follows:

step one, repeating the step two to the step four for a plurality of times to obtain a large number of alignment processing images as image samples, and manually determining the production serial number corresponding to each image sample;

step two, introducing a neural network module into python, extracting an image sample obtained in the step one, and identifying the production sequence number by using an MPL neural network algorithm to obtain an identification result of the image sample;

thirdly, judging whether the identification result is matched with the manually identified production sequence number result, and returning to the second step to identify the production sequence number of the image sample by reusing the MPL neural network algorithm if the identification result is failed to be matched with the manually identified production sequence number result; if the matching is successful, returning to the second step to extract the next image sample for identifying the production serial number until all the image samples can be successfully identified to obtain an image production serial number identification model;

step six, in the actual operation of the oil-gas module, selecting a steel structure, sequentially performing the processes of the step two, the step three and the step four, and calculating the alignment processing image obtained in the step four by using the image production sequence number prediction model obtained in the step five in python to obtain the production sequence number of the steel structure;

and step seven, the computer sends a request to the server through the network, finds the steel structure corresponding to the production sequence in the step six in the database of the server, and displays the basic information of the steel structure on a display screen of the operation site through network transmission, so that the site operation is facilitated.

Images