A method for extracting the contour and measuring the dimensions of forgings based on infrared thermal imaging

Abstract

This invention provides a method for extracting the contour and measuring the dimensions of forgings based on infrared thermal imaging, comprising the following steps: acquiring infrared thermal images of forgings in complex scenarios using infrared thermal imaging technology; reconstructing the low-resolution infrared thermal images of forgings using image super-resolution technology; removing the low-temperature background based on the temperature difference in the reconstructed infrared thermal images of forgings; extracting the contour of the forgings after background removal using edge contour extraction technology; establishing the circumscribed rectangle of the forging contour, obtaining the vertex coordinates of the circumscribed rectangle, and calculating the dimensional information of the forgings; and transmitting the forging contour and dimensional information to a host computer for display. The method for extracting the contour and measuring the dimensions of forgings based on infrared thermal imaging proposed in this invention can adapt to complex working environments such as high-temperature thermal radiation, strong background noise, and large forging dimensions, and improves the real-time performance, accuracy, and completeness of the measurement.

Description

Technical Field

[0001] This invention belongs to the field of online measurement of forgings, and specifically relates to a method for extracting the contour and measuring the dimensions of forgings based on infrared thermal imaging. Background Technology

[0002] Extracting the contour and measuring the dimensions of forgings in complex scenarios has long been a technical challenge that has been difficult to solve effectively both domestically and internationally. Early methods for measuring forgings in complex scenarios mainly focused on contact measurement methods such as calipers and measuring rods. These methods require manual operation and are subject to high-temperature environments, resulting in the inability to achieve accurate online measurement and low repeatability. At the same time, the high-temperature heat radiation environment poses safety hazards to operators.

[0003] To address the aforementioned issues, non-contact measurement methods have been introduced into the measurement of forging dimensions in complex scenarios in recent years. The main non-contact measurement methods for forgings in complex scenarios include laser scanning, CCD imaging, and structured light projection. The patent application submitted by Jia Zhenyuan et al. of Dalian University of Technology, entitled "A Non-Contact Measurement Method for Hot-State Dimensions of Forgings" (patent number CN 101713640 B), uses projected structured light strips combined with binocular vision technology to capture light stripe images of high-temperature forgings. By extracting the morphological feature points and lines of the forging, the forging dimensions are measured. However, this method only measures cylindrical and square forgings, does not consider end deformation, central bulging, etc., and cannot obtain complete contour information for forging areas without projected light strips. The patent application submitted by Liu Renqiang of China First Heavy Machinery Co., Ltd., entitled "Online Measurement Device and Method for Large Free Forgings" (patent number CN 103411530 B), uses a combination of projected laser beam method, infrared thermometry automatic edge finding method, laser ranging method, and CCD image measurement to obtain the diameter, length, and outline of the forging within the horizontal cross-section. However, the edge finding method does not consider the edge blurring caused by heat exchange between the forging and the anvil. Furthermore, fusing the information from the projected laser beam method, thermometry edge finding method, and laser ranging method requires a significant amount of time, which is insufficient for the practical requirements of free forging. All of these methods fail to consider the impact of complex working conditions during forging, such as forging movement, image blurring, downsampling, background noise, and measurement distance, on the real-time performance, accuracy, and completeness of the measurement.

[0004] Existing forging measurement methods suffer from poor real-time performance, low measurement accuracy, incomplete measurement data, short measurement distance, and difficulty in adapting to complex environments such as high-temperature heat radiation, strong background noise, and large forging dimensions. Therefore, there is an urgent need to design an online method for forging contour extraction and dimension measurement in complex scenarios that can adapt to the complex environment of industrial sites, achieve rapid and accurate measurement of important dimensions of forgings, and obtain complete contour information of forgings for traceability analysis. Summary of the Invention

[0005] The technical problem to be solved by this invention is to propose a method for extracting the outline and measuring the size of forgings based on infrared thermal imaging. The method acquires infrared thermal images of forgings using infrared thermal imaging technology, reconstructs the infrared thermal images of forgings at high resolution using image super-resolution technology, removes low-temperature backgrounds by utilizing temperature differences in the infrared thermal images, extracts the outline of the forgings after background removal using edge contour extraction technology, establishes the circumscribed rectangle of the forging outline, and calculates the size information of the forgings based on the vertex coordinates. This method can overcome the problems of poor real-time performance, low measurement accuracy, incomplete measurement data, and short measurement distance of previous measurement methods, and can adapt to complex scenarios such as high-temperature thermal radiation, strong background noise, and large forging size.

[0006] To achieve the above objectives, the present invention adopts the following technical solution:

[0007] S1: Obtain infrared thermal images of forgings in complex scenarios based on infrared thermal imaging technology;

[0008] S2: Reconstructing low-resolution infrared thermal images of forgings using image super-resolution technology;

[0009] S3: Remove low-temperature background based on temperature differences in the reconstructed infrared thermal image of the forging;

[0010] S4: The edge contour extraction technology is used to extract the contour of the forging after the background is removed;

[0011] S5: Establish the circumscribed rectangle of the forging outline, obtain the vertex coordinates of the circumscribed rectangle, and calculate the Euclidean distance between adjacent vertex coordinates to obtain the dimensional information of the forging;

[0012] S6: Transmit the forging outline and dimensional information to the host computer for display.

[0013] 2. Further, the specific steps of the infrared thermal image reconstruction method in step S2 are as follows:

[0014] S2.1: Extract low-resolution single-frame infrared thermal image of the forging Y LR ;

[0015] S2.2: Y LR The training samples are input into the training network for learning. The motion factor M(x) of the forging during the forging process, the imaging blur factor B(x) of the forging, and the downsampling factor S(x) of the thermal imager are learned. The degradation function H is used to represent the degradation process of the infrared thermal image of the forging, where H(x) = M(x) × B(x) × S(x).

[0016] S2.3: Obtain the infrared thermal image noise N(x) of the forging, and combine it with the degradation function to analyze the single-frame infrared thermal image Y of the forging. LR Perform reverse engineering to obtain the high-resolution image X.HR , where X HR = H -1 (x)(Y LR -N(x)).

[0017] 3. Further, the specific steps of the low-temperature background removal method in step S3 are as follows:

[0018] S3.1: Convert the reconstructed infrared thermal image of the forging into a temperature thermal image based on the hue / saturation / brightness color model;

[0019] S3.2: According to the temperature-color mapping model, traverse the color space difference value t between any two points m1 and m2 in the whole image, where t = ||m1 - m2||;

[0020] S3.3: Perform clustering analysis according to the color difference value, design a judgment threshold t0, when t < t0, it is judged as the same class, otherwise it is judged as different classes, and divide the forging image into two classes C1 and C2;

[0021] S3.4: Obtain the average pixels of C1 and C2, judge that the area with the larger average pixels is the high-temperature forging area, and the one with the smaller average pixels is the low-temperature background area. Set the pixels of the low-temperature background area with smaller average pixels to 0 to remove the low-temperature background.

[0022] The beneficial effects of the present invention are as follows:

[0023] A method for forging contour extraction and size measurement based on infrared thermal imaging is proposed to achieve high-precision extraction of the contour of large high-temperature forgings and rapid measurement of size; solve the problems of high-temperature thermal radiation and long-distance imaging in complex scenarios based on infrared thermal imaging technology; use image super-resolution technology to perform high-resolution reconstruction on the infrared thermal image of the forging, and solve the problem of reduced image resolution caused by motion changes, imaging blurring, downsampling and other noises during forging shooting; remove the low-temperature background based on the temperature difference of the infrared thermal image of the forging to obtain a forging image without background interference; obtain the forging contour after removing the background through edge contour extraction technology, establish the circumscribed rectangle of the forging contour, calculate the size information of the forging based on the vertex coordinates, and display the contour and size of the forging in real time, providing a new method for forging contour extraction and size measurement in complex scenarios. Brief Description of the Drawings

[0024] Figure 1 Flowchart of a method for forging contour extraction and size measurement based on infrared thermal imaging;

[0025] Figure 2 Original infrared thermal image of the forging;

[0026] Figure 3 Schematic diagram of super-resolution reconstruction of the infrared thermal image of the forging

[0027] Figure 4 Image showing the effect of removing the background from a forging;

[0028] Figure 5 Rendering of the forging outline;

[0029] Figure 6 Calculation results of forging dimensions. Detailed Implementation

[0030] The present invention will now be described in further detail with reference to the accompanying drawings and specific embodiments. The embodiments of the present invention are only a part, not all, of the embodiments. All other embodiments obtained by those skilled in the art without creative effort are within the protection scope of the present invention.

[0031] The following example uses a large high-temperature forging from a forging workshop as an example.

[0032] A flowchart of a method for extracting the contour and measuring the dimensions of forgings based on infrared thermal imaging is shown below. Figure 1 As shown, the measurement method can be divided into the following steps:

[0033] S1: Obtaining original infrared thermal images of forgings in complex scenarios based on infrared thermal imaging technology, such as... Figure 2 As shown;

[0034] S2: Image super-resolution technology is used to reconstruct the low-resolution infrared thermal image of the forging. A schematic diagram of the reconstruction is shown below. Figure 3 As shown;

[0035] S3: Removal of low-temperature background based on temperature differences in the reconstructed infrared thermal image of the forging; removal effect. Figure 4 As shown;

[0036] S4: Edge contour extraction technology is used to extract the contour of the forging after removing the background. The extraction effect is as follows: Figure 5 As shown;

[0037] S5: Establish the circumscribed rectangle of the forging profile, obtain the vertex coordinates of the circumscribed rectangle, and calculate the Euclidean distance between adjacent vertex coordinates to obtain the dimensional information of the forging. The calculation result is as follows: Figure 6 As shown;

[0038] S6: Transmit the forging outline and dimensional information to the host computer for display.

[0039] Preferably, in step S1, the infrared thermal imaging technology uses an infrared thermal imager to photograph the forging, and the sampling frequency and shooting time are set using a controller.

[0040] Preferably, the infrared thermal image reconstruction method in step S2 specifically comprises the following steps:

[0041] S2.1: Extract the single-frame infrared thermal image Y of the forging at low resolution LR ;

[0042] S2.2: Use Y LR as a training sample and input it into a deep learning neural network for learning. Learn the motion factor M(x) of the forging, the imaging blur factor B(x) of the forging, and the downsampling factor S(x) of the thermal imager. Use the degradation function H to represent the degradation process of the infrared thermal image of the forging, where H(x) = M(x) × B(x) × S(x);

[0043] S2.3: Obtain the infrared thermal image noise N(x) of the forging. Combine the degradation function to perform inverse solution on the single-frame infrared thermal image Y of the forging LR to obtain the high-resolution image X HR , where X HR = H -1 (x)(Y LR - N(x)).

[0044] Preferably, the specific steps of the low-temperature background removal method in step S3 are as follows:

[0045] S3.1: Use the hsv operator to convert the reconstructed infrared thermal image of the forging into a temperature thermal image based on the hue / saturation / brightness color model;

[0046] S3.2: According to the temperature-color mapping model, traverse the color space difference value t between any two points m1 and m2 in the whole image, where t = ||m1 - m2||;

[0047] S3.3: Perform clustering analysis according to the color difference value. Design the judgment threshold t0 to be 220. When t < t0, it is judged as the same class, otherwise it is judged as different classes. Divide the forging image into two classes C1 and C2;

[0048] S3.4: Obtain the average pixels of C1 and C2. Judge that the area with the larger average pixel is the high-temperature forging area, and the area with the smaller average pixel is the low-temperature background area. Set the pixels of the low-temperature background area with the smaller average pixel to 0 to remove the low-temperature background.

[0049] Preferably, the edge contour extraction technology in step S4 uses the Canny operator to find the edge contour of the forging, and its parameter settings are Canny(img, 70, 200).

[0050] Preferably, the calculation method in step S5 obtains the coordinates of the vertices of the circumscribed rectangle of the forging edge contour A(x1, y1), B(x2, y2), C(x3, y3). The Euclidean distance between points A and B represents the length L of the forging, where The Euclidean distance between points A and C represents the height H of the forging, where

[0051] Examples of the present invention have been described above with reference to the accompanying drawings. However, the present invention is not limited to the specific embodiments described above. The specific embodiments described above are merely exemplary and not limiting. Any invention that does not exceed the claims of the present invention is within the protection scope of the present invention.

Claims

1. A method for forging profile extraction and dimension measurement based on infrared thermography, characterized in that The forging profile extraction and size measurement are realized based on infrared thermal image reconstruction and infrared thermal image temperature difference. The specific steps of the method are as follows: S1: obtaining the forging infrared thermal image under a complex scene based on infrared thermal imaging technology; S2: using image super-resolution technology to reconstruct the low-resolution forging infrared thermal image, first extract the low-resolution forging infrared thermal image Y LR As a training sample input to the training network learning, learning the movement factor M(x) of the forging, the imaging blur factor B(x) of the forging, the down-sampling factor S(x) of the infrared thermal imager, and using the degradation function H(x)=M(x)×B(x)×S(x) to represent the above degradation process, then obtain the infrared image noise N(x), combine the degradation function Y LR To solve inversely, obtain the high-resolution image X HR , which can be expressed as X HR =H -1 (x)(Y LR -N(x)) S3: removing the low-temperature background based on the temperature difference of the reconstructed forging infrared thermal image. Firstly, the reconstructed forging infrared thermal image is converted into a temperature thermal image based on a hue / saturation / lightness color model. Then, the color space difference value t between any two points m1 and m2 in the image is traversed, wherein t=||m1-m2||. Then, a judgment threshold t0 is designed. When t<t0, it is judged as the same class, otherwise, it is judged as different classes. The forging image can be divided into two classes C1 and C2. The average pixels of C1 and C2 are obtained. The region with large average pixel is judged as the high-temperature forging region, and the region with small average pixel is judged as the low-temperature background region. The pixels of the low-temperature background region with small average pixel are set to 0, and then the low-temperature background is removed; S4: extracting the profile of the forging after removing the background by using the edge profile extraction technology; S5: establishing the circumscribed rectangle of the forging profile, obtaining the vertex coordinates of the circumscribed rectangle, and calculating the Euclidean distance of the adjacent vertex coordinates to obtain the size information of the forging; S6: transmitting the forging profile and size information to the host computer for display.

Images