CN106290379A - Rail surface defects based on Surface scan camera detection device and method - Google Patents

Abstract

The invention discloses a rail surface defect detection device and method based on an area scanning camera, which can improve the efficiency and accuracy of detection and greatly reduce the workload of workers. The device includes: a surface scanning camera, a data acquisition card and a processing module; wherein, the surface scanning camera is used to collect the surface image of the rail to be detected; the data collection card is used to obtain the image, and the The image is sent to the processing module; the processing module is used to process the image to obtain a defect feature value, and input the defect feature value into a pre-trained classifier to obtain the rail to be detected The category the defect belongs to.

Description

Rail surface defect detection device and method based on area scanning camera

technical field

The invention relates to the field of rail surface defect detection, in particular to a rail surface defect detection device and method based on an area scanning camera.

Background technique

With the development of the railway industry, high-speed rail has an indispensable position in transportation. However, there will be certain defects in the rails during railway operation or after processing, which will have a direct impact on the service life and mechanical properties of the materials, and will gradually develop and deteriorate, eventually leading to material failure and fracture, and may also cause Great loss of life or property. Now with the introduction of advanced production technology, such as high-carbon steel and steel clean production process, the probability of internal defects of rails is decreasing day by day, and it is more and more common for rails to break due to surface defects of rails; surface defects of rails can also cause wheels Abrasion increases, resulting in peeling off blocks, making derailment more likely to occur, so the detection of rail surface defects is particularly important.

In terms of rail defect detection, the more mature and widely used technologies are automatic and semi-automatic methods such as ultrasonic testing and pulsed eddy current testing. However, on the one hand, these methods have high detection costs and complicated equipment, and on the other hand, it is difficult to distinguish internal defects and surface defects of the rail, requiring manual review with small instruments.

Contents of the invention

Aiming at the deficiencies and defects of the prior art, the present invention provides a rail surface defect detection device and method based on an area scanning camera.

On the one hand, an embodiment of the present invention proposes a rail surface defect detection device based on an area scanning camera, including:

area scan camera, data acquisition card and processing module; where,

The surface scanning camera is used to collect the surface image of the rail to be detected;

The data acquisition card is used to acquire the image and send the image to the processing module;

The processing module is configured to process the image to obtain defect feature values, and input the defect feature values into a pre-trained classifier to obtain the category to which the rail defects to be detected belong.

On the other hand, an embodiment of the present invention proposes a rail surface defect detection method, including:

Using the surface scanning camera to collect a surface image of the rail to be detected;

using the data acquisition card to acquire the image, and sending the image to the processing module;

The image is processed by the processing module to obtain defect feature values, and the defect feature values are input into a pre-trained classifier to obtain the categories to which the defects of the rail to be detected belong.

The rail surface defect detection device and method based on the surface scanning camera provided by the embodiments of the present invention uses the processing module to detect the category of the defect of the surface image of the rail to be detected collected by the surface scanning camera through the classifier, which can well realize the detection of rail defects. The identification and classification of defects improves the efficiency and accuracy of detection, and greatly reduces the workload of workers. At the same time, the invention overcomes the shortcomings of existing automatic and semi-automatic methods such as ultrasonic testing and pulsed eddy current testing, which have high costs and complex equipment. , under the premise of ensuring the detection efficiency and accuracy, the identification and classification of rail surface defects are realized.

Description of drawings

Fig. 1 is a structural schematic diagram of an embodiment of a rail surface defect detection device based on an area scanning camera in the present invention;

Fig. 2 is a schematic diagram of the specific process involved in an embodiment of the surface-scanning camera-based rail surface defect detection device of the present invention;

Fig. 3 is a structural schematic diagram of the camera mobile platform of the present invention;

Fig. 4 is a detailed view of the guide rail part of the camera mobile platform described in Fig. 3;

Fig. 5 is a schematic diagram of an overall process of an embodiment of a method for detecting a surface defect of a rail according to the present invention.

detailed description

In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly described below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are the Some, but not all, embodiments are invented. Based on the embodiments of the present invention, all other embodiments obtained by persons of ordinary skill in the art without making creative efforts belong to the protection scope of the present invention.

Referring to Fig. 1, this embodiment discloses a rail surface defect detection device based on an area scanning camera, including:

Surface scan camera 11, data acquisition card 12 and processing module 13; Wherein,

The surface scanning camera 11 is used to collect surface images of rails to be detected;

In practical applications, there will be distortion in the camera lens, which is a phenomenon of distortion and distortion of the image caused by the different magnification ratios of different areas on the focal plane. Distortion will affect the accuracy of defect detection, and calibrating the camera can achieve the effect of eliminating distortion. Therefore, before using the camera to collect the surface image of the rail to be inspected, it needs to be calibrated first, as shown in Figure 2.

The process of camera calibration is actually the process of solving 12 internal and external parameters of the camera. One of the tools needed for calibration is the calibration board. Commonly used calibration boards include solid circle array pattern and chess array pattern. The size can be selected according to the size of the field of view.

After the camera position and attitude are fixed, several images of the calibration board are collected. These images need to make the calibration boards in different positions cover the entire camera field of view, and at the same time, the calibration boards should have as many poses as possible. Input the image of the calibration board into the software program written in advance with the machine vision software HALCON, run it to get the internal and external parameters of the camera, and use the internal and external parameters of the camera to calibrate the camera.

In this embodiment, the camera can use an area array CMOS camera ARTCAM-300MI-WOM, which can capture images with a maximum of 2048×1536 pixels. Such a high resolution can ensure that the collected rail defect images can save as much detailed information as possible. This ensures that as few defects as possible are missed. And because each row of pixels can reach a maximum of about 6000, therefore, using this camera to acquire images, each millimeter can include about 20 pixels, combined with the subsequent image processing algorithm, can ensure that the results of defect position detection meet the accuracy requirements. Since the field of view of the camera is about 15cm×10cm during the experiment, the detection result of surface defects can be accurate to 0.1mm by using this camera.

The data acquisition card 12 is configured to acquire the image, and send the image to the processing module 13;

The processing module 13 is configured to process the image to obtain a defect feature value, and input the defect feature value into a pre-trained classifier (which may be a Gaussian classifier) to obtain the defect of the rail to be detected category to which it belongs.

The processing module 13 can be set in a PC. As shown in Figure 2, the processing module 13 needs to train the classifier before processing the surface image of the rail to be detected, and the specific process is as follows:

Perform preprocessing and feature extraction on the surface image of the rail sample of the category of the known defect to obtain the defect feature value of the rail sample, and use the category of the defect of the rail sample and the defect feature of the rail sample value to train the classifier on.

As shown in Figure 2, after the classifier is trained, the preprocessing submodule of the processing module 13 will preprocess the surface image of the rail to be detected, and the calculation submodule will perform feature extraction on the preprocessed image (that is, obtain the defect information) to get the defect feature value. Wherein, the preprocessing includes: sequentially performing filter filtering, image enhancement and rail area positioning processing on the image. It should be noted that the filter filtering mainly uses the Gaussian function (mean_image function) to denoise and smooth the image, which is very effective for removing noise that obeys a normal distribution.

The feature extraction includes: performing image grayscale on the rail region obtained by the rail region positioning processing; performing image segmentation on the result of the grayscale image; performing morphological processing on the result of the image segmentation; Defect calibration is performed on the results of scientific processing; the characteristic value of the defect is calculated to obtain the characteristic value of the defect. Image grayscale is mainly to highlight the characteristics of the image. Specifically, the decompose3 operator can be used to convert the color image into a black and white image, which is convenient for subsequent effective processing. Image segmentation mainly extracts defects based on the differences in gray values, which is conducive to convenient, quick and effective extraction of defect information. It has the characteristics of simple implementation, small amount of calculation, and relatively stable performance. It is especially suitable for images where the target and background occupy different gray levels. . Morphological processing includes erosion and dilation operations, and corrosion dilation is the basis of morphological processing. Through the corrosion and expansion operation, unwanted details can be ignored, the characteristics of defects can be enhanced, the effect of enhancing the picture can be achieved, and the area of interest can be screened out. For defect calibration, the canny operator can be used to select and locate the edge of the defect image. The feature values extracted by the feature extraction may include the length and position information of the defect.

It should be noted that the preprocessing and feature extraction of the surface image of the rail sample of the category to which the known defect belongs is the same as the corresponding processing process of the surface image of the rail to be detected, and will not be repeated here.

The defect categories that can be identified by using the embodiment of the present invention may include peeled off blocks, fish scale peeling, cracks and rust, and of course other types of defect information may also be added to the classifier for defect identification.

The rail surface defect detection device based on the surface scanning camera provided by the embodiment of the present invention uses the processing module to detect the category of the defect of the surface image of the rail to be detected collected by the surface scanning camera through the classifier, which can well realize the detection of rail defects Identify and classify, improve the efficiency and accuracy of detection. At the same time, the present invention overcomes the shortcomings of existing automatic and semi-automatic methods such as ultrasonic detection and pulsed eddy current detection, such as high cost and complicated equipment. On the premise of ensuring detection efficiency and accuracy Under this method, the identification and classification of rail surface defects are realized.

On the basis of the foregoing device embodiments, the device may also include:

motor and camera movement platform; where,

The area scan camera is set on the camera mobile platform,

The motor is connected to the camera moving platform for receiving control instructions from the processing module, and drives the camera moving platform to move through the transmission mechanism according to the control instructions, thereby driving the area scanning camera to move in three-dimensional space .

As shown in Figure 3, it is a schematic structural view of the camera mobile platform of the present invention. The camera mobile platform is mainly used to control the movement of the camera. The platform is respectively provided with movable sliders along the three directions of XYZ, which are used to drive the camera in three-dimensional space. sports. In Fig. 3, 1, 3 and 4 are sliders, 2 is a rail to be detected, and 5 is a surface scanning camera. The slider at position 1 is used to control the forward and backward movement of the camera, and the sliders at position 3 and 4 are used to control the up and down and left and right movements of the camera respectively, so as to realize the movement of the camera in three-dimensional space. The details of the guide rail are shown in Figure 4 . The surface scanning camera is controlled to move in three-dimensional space through the camera moving platform to scan and obtain the surface image of the rail to be detected.

As shown in Figure 3, the information transmission between the processing module and the data acquisition card and between the data acquisition card and the camera is bidirectional. The processing module sends a control signal to realize the motion control of the camera through the digital acquisition card, and the image information of the defect collected by the camera is transmitted to the processing module through the digital acquisition card.

It should be noted that the overall structure of the camera mobile platform shown in Figure 3 and Figure 4 adopts a cuboid structure, but this structure does not constitute a limitation to the present invention, it is only an embodiment of the present invention, professionals in the field can build any controllable A platform on which the camera moves in 3D space.

The software compilation environment of the device mainly uses the interactive programming interface HDevelop in HALCON to carry out the software design of the whole device. The software program can be directly embedded into the hardware of the device to identify and classify defects directly, which is convenient, efficient and low cost.

Referring to Figure 5, this embodiment discloses a method for detecting surface defects of rails based on an area scanning camera, including:

S1. Using the surface scanning camera to collect the surface image of the rail to be detected;

S2. Use the data acquisition card to acquire the image, and send the image to the processing module;

S3. Using the processing module to process the image to obtain defect feature values, and input the defect feature values into a pre-trained classifier to obtain the categories to which the defects of the rail to be detected belong.

The rail surface defect detection method based on the surface scanning camera provided by the embodiment of the present invention uses the processing module to detect the category of the defect of the surface image of the rail to be detected collected by the surface scanning camera through the classifier, which can well realize the detection of rail defects Identify and classify, improve the efficiency and accuracy of detection. At the same time, the present invention overcomes the shortcomings of existing automatic and semi-automatic methods such as ultrasonic detection and pulsed eddy current detection, such as high cost and complicated equipment. On the premise of ensuring detection efficiency and accuracy Under this method, the identification and classification of rail surface defects are realized.

Although the embodiments of the present invention have been described in conjunction with the accompanying drawings, those skilled in the art can make various modifications and variations without departing from the spirit and scope of the present invention. within the bounds of the requirements.

Claims (9)

1. A rail surface defect detection device based on a surface scanning camera is characterized by comprising:

the system comprises a surface scanning camera, a data acquisition card and a processing module; wherein,

the surface scanning camera is used for acquiring a surface image of the steel rail to be detected;

the data acquisition card is used for acquiring the image and sending the image to the processing module;

and the processing module is used for processing the image to obtain a defect characteristic value, and inputting the defect characteristic value into a pre-trained classifier to obtain the category to which the defect of the steel rail to be detected belongs.

2. The apparatus of claim 1, wherein the processing module comprises:

the preprocessing submodule is used for preprocessing the image;

and the calculation submodule is used for extracting the defects of the image processed by the preprocessing submodule and calculating the characteristic values of the extracted defects to obtain the defect characteristic values.

3. The apparatus of claim 2, wherein the pre-processing submodule is specifically configured to:

and sequentially carrying out filter filtering, image enhancement and steel rail positioning processing on the image.

4. The apparatus according to claim 3, wherein the computation submodule is specifically configured to:

carrying out image graying on the steel rail area obtained by the steel rail positioning processing;

performing image segmentation on the image graying result;

performing morphological processing on the result of the image segmentation;

performing defect calibration on the result of the morphological processing;

and calculating the obtained characteristic value of the defect to obtain the defect characteristic value.

5. The apparatus according to claim 1, wherein the processing module is further configured to, before processing the image to obtain the defect feature value, process a surface image of a steel rail sample of a category to which the defect belongs to obtain the defect feature value of the steel rail sample, and train the classifier using the category to which the defect of the steel rail sample belongs and the defect feature value of the steel rail sample.

6. The apparatus of claim 1, further comprising:

a motor and a camera moving platform; wherein,

the area scanning camera is arranged on the camera moving platform,

the motor is connected with the camera moving platform and used for receiving the control instruction from the processing module and driving the camera moving platform to move through the transmission mechanism according to the control instruction, so that the surface scanning camera is driven to move in a three-dimensional space.

7. The apparatus of claim 1,

the surface scanning camera is also used for acquiring images of the calibration plates with different postures which are paved in the camera view field before acquiring the surface image of the steel rail to be detected;

the processing module is used for acquiring the image of the calibration board, calculating the internal and external parameters of the surface scanning camera according to the image of the calibration board, and calibrating the surface scanning camera by using the internal and external parameters.

8. The apparatus of claim 1, wherein the area scan camera is an area array CMOS camera ARTCAM-300 MI-WOM.

9. A rail surface defect detection method based on the device of claim 1, which is characterized by comprising the following steps:

acquiring a surface image of the steel rail to be detected by using the surface scanning camera;

acquiring the image by using the data acquisition card, and sending the image to the processing module;

and processing the image by using the processing module to obtain a defect characteristic value, and inputting the defect characteristic value into a pre-trained classifier to obtain the category to which the defect of the steel rail to be detected belongs.