KR102449124B1 - System for concrete crack detection using 2D and 3D image - Google Patents

Abstract

The present invention relates to a concrete crack detection system using 2D and 3D configurations, and to a concrete crack detection system using 2D and 3D configurations to recognize and detect cracks through image analysis. Through 2D image and 3D shape data analysis, when a crack occurs in the upper part of concrete, the depth information is used to solve the false detection of recognizing foreign substances or other error conditions in the image as a crack, and confirm the false detection by foreign substances in the 2D image crack analysis process In order to do this, it has the effect of verifying the analysis of the crack location using the height information in the 3D inspection.

Description

Concrete crack detection system using 2D and 3D configuration {System for concrete crack detection using 2D and 3D image}

The present invention relates to a concrete crack detection system using a 2D and 3D configuration, and more particularly, to a concrete crack detection system using a 2D and 3D configuration to recognize a crack through image analysis on a concrete ballast and detect a crack width. it's about

The general machine vision inspection system consists of a camera and lighting (or line laser), and the typical configuration methods are divided into 2D inspection equipment and 3D inspection equipment. 2D inspection uses a camera and light to acquire an image and then analyzes the defect using a brightness value or a specific pattern within the image, and 3D inspection uses a camera and line laser to acquire shape data for the image, and height information to check the numerical value of the shape.

Each technology has advantages and disadvantages, and the 2D inspection is a crack measurement method that can quickly and precisely measure the crack propagation length and width of the surface in real time in the prior art Patent Publication No. 2009-0126362, which measures cracks using a camera and lighting. , there is a problem in that when a crack occurs in the upper part of concrete, the depth information cannot be known, so foreign substances or other error conditions in the image can be recognized as cracks. In addition, in the case of a 3D inspection that measures cracks using a laser image, the reliability of cracks using depth information is better than that of the 2D method, but it has advantages and disadvantages in that it is acquired as 3D shape data rather than an image form that can be seen with the naked eye.

The present invention is to solve the above problems, and through 2D image and 3D shape data analysis, using depth information when cracks occur in the upper part of concrete, 2D and An object of the present invention is to provide a concrete crack detection system using a 3D configuration.

The present invention is a concrete crack detection system using a 2D and 3D configuration, in which the top and bottom of the image sensor are configured as 2D and 3D areas, respectively, with one camera, and the upper area and the lower area are for acquiring images using a 2D lighting device. an inspection device comprising a region for acquiring shape information using a 3D line laser; 2D analysis of performing crack analysis on a concrete ballast on a 2D image to collect image data of the inspection device and generating one image frame, and filtering and crack analysis using depth information for crack locations in 3D shape data and a collection device having a data collection and analysis device to verify; and a tachometer installed on the wheel axle of a railroad vehicle to generate a signal according to the vehicle moving distance to synchronize the distance to the inspection device through the collecting device.

Preferably, the inspection device is composed of a 2D image sensor (CCD or CMOS), which acquires data in 2D and 3D according to the lighting configuration, and uses an optical filter to filter out interference between individual lights; an FPGA module that receives image data from the image sensor of the camera device and stores the data classified into 2D and 3D areas in a separate memory space to separately classify 2D image and 3D shape data and transmit the data to the collecting device; A 2D lighting device consisting of a line-type lighting as a lighting device in the visible light region and consisting of an LED or a laser; and a 3D laser device configured as a line-type laser as an illumination device in the infrared region.

Preferably, the collection device includes: an interface device that collects high-speed image data received from the inspection device, generates one image frame, and transmits the generated image frame data to the data collection and analysis device; a data collection and analysis device that analyzes and verifies cracks by filtering 2D crack inspection results using 2D analysis for performing crack analysis on the concrete ballast in 2D images and depth information for crack locations in 3D shape data; and a distance synchronization device that analyzes the moving distance information of the vehicle received from the tachometer and generates a shooting trigger signal at equal intervals to the inspection device.

Preferably, the data collection and analysis device includes: a data acquisition module that receives the original concrete image taken by the image sensor of the camera device as an image input and classifies 2D and 3D data into 2D images and 3D shape images, respectively; a noise removal module that separates the structures and rail tracks of the concrete ballast, which are linear patterns in 2D image and 3D shape data, from the ballast because cracks on the concrete ballast are created on the top of the ballast for ballast recognition; a filtering module that performs a process for removing a concrete structure and a rail track from a map image using a straight line detection algorithm; In the crack analysis of 2D images, the standard pattern of cracks is input in advance, and the process of recognizing cracks that match the standard pattern of cracks in the acquired concrete ballast image, and depth information analysis data deeper than the preset depth based on the height information of the upper part of the ballast a crack analysis module that performs the process of extracting ; and a crack analysis verification module that verifies the analysis of the crack location by using height information in the 3D inspection in order to confirm erroneous detection by foreign substances in the 2D image crack analysis process of the crack analysis module.

Preferably, the inspection device classifies the region of interest in which only a portion of the image sensor is acquired into the 2D and 3D regions, and the imaging is acquired at equal intervals according to the moving distance of the vehicle, and the tachometer installed on the wheel is the vehicle's Position information synchronization is performed between the 2D and 3D data using an equal interval signal by measuring the moving distance.

According to the present invention, through 2D image and 3D shape data analysis, when a crack occurs in the upper part of concrete, erroneous detection of recognizing a foreign substance or other error condition in an image as a crack is solved by using depth information, and in the 2D image crack analysis process, it is It has the effect of verifying the analysis of the crack location by using the height information in the 3D inspection to confirm the false detection by the

1 is a view showing a schematic configuration of a concrete crack detection system using a 2D and 3D configuration according to an embodiment of the present invention.
2 is a view showing a camera and lighting configuration of a concrete crack detection system using a 2D and 3D configuration according to an embodiment of the present invention.
3 is a view showing the acquisition and transmission of 2D and 3D data using the FPGA module of the concrete crack detection system using the 2D and 3D configuration according to an embodiment of the present invention.

Hereinafter, the present invention will be described in more detail with reference to the drawings. It should be noted that the same components in the drawings are denoted by the same reference numerals wherever possible. In addition, descriptions of well-known functions and configurations that may unnecessarily obscure the gist of the present invention will be omitted.

First, according to an embodiment of the present invention, 2D image collection for user's visual inspection after crack analysis and system simplification for concrete defect inspection with a configuration that acquires 2D and 3D data with one camera instead of multiple cameras Includes configuration for 3D data inspection to increase inspection accuracy.

An embodiment of the present invention will be described in detail with reference to the drawings. 1 is a view showing a schematic configuration of a concrete crack detection system using a 2D and 3D configuration according to an embodiment of the present invention. As shown in FIG. 1 , the concrete crack detection system using 2D and 3D configurations according to an embodiment of the present invention includes an inspection device 100 , a collecting device 200 , and a tachometer 300 in a large view.

Inspection device 100 is a camera device 110, FPGA module 120, 2D lighting device 130, 3D laser device 140 in a configuration that acquires 2D and 3D data with one camera instead of a plurality of cameras include

The camera device 110 is composed of a 2D image sensor (CCD or CMOS) to obtain 2D and 3D data according to the lighting configuration, and filters the interference between individual lights by using an optical filter. In addition, the camera device 110 uses a region of interest (ROI) function for high-speed image acquisition, and the ROI is acquired by dividing it into 2D and 3D regions. The FPGA module 120 receives image data from the image sensor of the camera device, stores the data classified into 2D and 3D areas in a separate memory space, classifies the 2D image and 3D shape data individually, and transmits the data to the collecting device do. Such an FPGA module may be included in a data acquisition and analysis device in a camera device or as an image input device (composed of FPGA) called a frame grabber according to a configuration method. The 2D lighting device 130 is a lighting device in the visible light region and is configured with line-type lighting and may be configured with LEDs or lasers. The 3D laser device 140 is an infrared region illumination device and is configured as a line-shaped laser.

The collection device 200 is configured to collect high-speed image data and generate one image frame, and includes an interface device 210 , a data collection and analysis device 220 , a distance synchronization device 230 , and a power supply 240 . include

The interface device 210 collects high-speed image data received from the inspection device and generates one image frame. Each image frame data generated by such an interface device is transmitted to a data collection and analysis device. The data collection and analysis device 220 is a data acquisition module ( 221 ), a noise removal module 222 , a filtering module 223 , a crack analysis module 224 , and a crack analysis verification module 225 . The detailed configuration of the data collection and analysis device 220 will be described in detail with reference to FIG. 4 below. The distance synchronization device 230 is a device that analyzes the moving distance information of the vehicle received from the tachometer and generates a shooting trigger signal at equal intervals to the inspection device. The power supply unit 240 is a device for supplying stable power to the collection and inspection device, and consists of a UPS.

The tachometer 300 is installed on the wheel axle of a railway vehicle to generate a signal according to the vehicle moving distance.

2 is a view showing a camera and lighting configuration of a concrete crack detection system using a 2D and 3D configuration according to an embodiment of the present invention. As shown in FIG. 2 , in the configuration of the photographing area of the camera device 110 , the upper and lower portions of the image sensor are configured as 2D and 3D areas, respectively, and the upper portion is an area for acquiring an image using a 2D lighting device, and the lower portion is 3D It consists of an area|region for which shape information is acquired using a line laser.

At this time, the position of the lighting for each component is configured so that the image is acquired with the lighting device perpendicular to the 2D area from the camera standard, and the 3D area is set at a specific angle based on the camera device to acquire the shape according to the height of the line laser. of light source to be irradiated.

In addition, the interference cancellation of illumination using an optical filter will be described as follows. An optical filter is applied as a band-pass filter that passes only a specific wavelength and blocks other wavelengths. It is a device that optically filters so that there is no interference when the light source of 2D light and 3D laser composed of visible light and infrared light is focused on the camera image sensor.

These optical filters are composed of an IR cut filter at the top and a visible light cut filter at the bottom based on the central part, so that when the light coming from each shooting area is focused on the image sensor, the light source is optically classified so that the corresponding light source is incident on the sensor.

3 is a view showing the acquisition and transmission of 2D and 3D data using the FPGA module of the concrete crack detection system using the 2D and 3D configuration according to an embodiment of the present invention. With reference to FIG. 3 , the separation of the ROI (Region of Interest) region of the sensor and the synchronization of position information between 2D and 3D data using equally spaced signals will be described as follows.

First, the separation of the ROI region of the sensor is a structure in which only a portion of the image sensor area is acquired in order to acquire high-speed data, and is classified into 2D and 3D regions using the ROI function. For the ROI area of the 2D inspection, one or two lines in the vertical direction designate the ROI, and the image is acquired for the corresponding section at each shooting point. In addition, the FPGA module stores the image data received as 2D line images in one memory area, and the FPGA module transmits a concrete image of one frame to the data collection and analysis device when data storage is completed in the corresponding memory. The ROI area of the 3D inspection is also structured in the same structure as the 2D inspection method, and when the data storage in the FPGA module is completed, the shape information of the concrete image of one frame is transmitted to the data collection and analysis device.

The synchronization of position information between 2D and 3D data using equal interval signals will be described as follows. Images are taken at equal intervals according to the moving distance of the vehicle, and the tachometer installed on the wheel measures the moving distance of the vehicle and transmits it to the distance synchronizer. The distance synchronizer analyzes the moving distance of the vehicle and generates a 0.5mm or equal interval shooting signal at a specific interval to the camera device (image sensor). . When shooting a concrete map with this signal configuration, the 2D and 3D frames are acquired the same, and if an offset is applied as much as the distance between the light and the laser (the distance between the 2D and 3D shooting) in the vertical direction of the shooting area (vehicle traveling direction), the concrete 2D image and 3D shape data at the same position are acquired with respect to the image acquisition data.

4 is an exemplary view showing the detailed configuration of the data collection and analysis apparatus of the concrete crack detection system using the 2D and 3D configuration according to an embodiment of the present invention. As shown in FIG. 4 , the data collection and analysis device 220 includes a data acquisition module 221 , a noise removal module 222 , a filtering module 223 , a crack analysis module 224 , and a crack analysis verification module 225 . includes

The data acquisition module 221 receives the original concrete image captured by the image sensor of the camera device as an image input, classifies 2D and 3D data into 2D images and 3D shape images, respectively, and stores them.

The noise removal module 222 separates the tracks such as structures and rails on the concrete ballast, which are linear patterns in the 2D image and 3D shape data, from the ballast because cracks on the concrete ballast are created on the upper part of the ballast for ballast recognition. go through

The filtering module 223 performs a process for removing concrete structures and rails from the map image using a straight line detection algorithm. In this embodiment, concrete structures and rails are removed from the image by using a linear detection algorithm using Hough transform, which is one of the techniques of linear detection in the image. Here, since the linear detection algorithm using the Hough transform is a known technique for linear detection in an image, a detailed description thereof will be omitted.

The crack analysis module 224 analyzes the crack depth information by analyzing the crack through the crack recognition process through crack analysis of the 2D image and the 3D formation inspection. This crack analysis module performs the process of pre-inputting the standard pattern of cracks in crack analysis of 2D images, and recognizing cracks that match the standard pattern of cracks in the acquired concrete image. In addition, the 3D shape inspection performs a process of extracting depth information analysis data deeper than a preset depth based on the height information of the upper part of the image.

The crack analysis verification module 225 performs a process of verifying the possibility of erroneous detection by other components in the 2D image crack analysis process of the crack analysis module 224 and the analysis of the crack location in the 3D shape inspection. Here, the crack is generated by cracking the concrete ballast, and the crack has height information. For this reason, the analysis verification uses the height information performed in the 3D inspection to reconfirm the possibility of erroneous detection by foreign substances (foreign substances such as paint, iron, branches, etc.) and other components for the crack position analyzed in the 2D image inspection. The process of verifying the analysis is performed.

5 is an exemplary view showing a crack measurement process using 2D/3D data of a concrete crack detection system using a 2D and 3D configuration according to an embodiment of the present invention. As shown in FIG. 5 , when the data collection and analysis apparatus acquires an image, the 2D and 3D data are classified into image data and shape data, respectively, and stored (a2, b2). Here, 2D data is image (2D) data and 3D data is shape (3D) data. Next, noise removal and filtering are performed in order to recognize an image from the classified image data and shape data (a3, b3). At this time, since the cracks on the concrete ballast are created on the top of the ballast, the ballast recognition separates the tracks such as structures and rails on the concrete ballast, which are linear patterns from the 2D image and 3D shape data, from the ballast, and detects the concrete structures and rails with a straight line detection algorithm. is used to perform the process for removing from the image. After image acquisition (a1), in the process (a4) of crack recognition by separating and acquiring image data (a2), crack recognition of 2D image is performed by inputting a standard pattern of cracks in advance, and The process of recognizing cracks consistent with the standard pattern is carried out. In addition, after acquiring shape data (b2) and performing processing for image recognition, in the process of analyzing depth information (b4), data deeper than the preset depth based on the height information of the upper part of the image is extracted through 3D shape inspection Analyze the depth information by performing the process. Previously, 2D image crack analysis and crack inspection through depth information analysis through 3D shape inspection were compared (c), and the possibility of erroneous detection due to other components in the 2D image crack analysis process and the location of cracks in 3D shape inspection were analyzed. Perform crack analysis verification (d) to verify

6 is an exemplary view showing the final output screen on which the defect determination of the concrete crack detection system using the 2D and 3D configuration according to an embodiment of the present invention.

Although described and illustrated in relation to a preferred embodiment for illustrating the technical idea of the present invention, the present invention is not limited to the configuration and operation as shown and described as such, and deviates from the scope of the technical idea. It will be apparent to those skilled in the art that many changes and modifications may be made to the present invention without reference to the invention. Accordingly, all such suitable alterations and modifications and equivalents are to be considered as falling within the scope of the present invention.

100: inspection device
110: camera device
120: FPGA module
130: 2D lighting device
140: 3D laser device
200: collection device
210: interface device
220: data collection and analysis device
230: distance synchronization device
240: power supply
300: tachometer

Claims (5)

Inspection consisting of the top and bottom of the image sensor with a single camera as 2D and 3D areas, respectively, and the upper area is an area that acquires images using a 2D lighting device, and the lower area is an area that obtains shape information using a 3D line laser Device;
2D analysis of performing crack analysis on a concrete ballast on a 2D image to collect image data of the inspection device and generating one image frame, and filtering and crack analysis using depth information for crack locations in 3D shape data and a collection device having a data collection and analysis device to verify; and
It includes; a tachometer installed on the wheel axle of a railroad vehicle to generate a signal according to the moving distance of the vehicle to synchronize the distance to the inspection device through the collecting device;
The data collection and analysis device receives the original concrete image taken from the image sensor of the camera device as an image input, classifies 2D and 3D data into 2D image and 3D shape image, respectively, and cracks in the concrete ballast for image recognition Since it is created on the top of the ballast, the structures and rail tracks of the concrete ballast, which are linear patterns in the 2D image and 3D shape image, are separated from the ballast, the concrete structures and rail tracks are filtered and removed from the ballast image, and crack analysis of the 2D image In the process of inputting the standard pattern of cracks in advance and recognizing cracks that match the standard pattern of cracks in the acquired concrete image, and extracting depth information analysis data deeper than the preset depth based on the height information of the upper part of the image A concrete crack detection system using a 2D and 3D configuration, characterized in that it verifies the analysis of the crack location using height information in the 3D inspection in order to confirm erroneous detection by foreign substances in the 2D image crack analysis process. According to claim 1,
The inspection device is
a camera device that is composed of a 2D image sensor (CCD or CMOS), acquires data in 2D and 3D depending on the lighting configuration, and filters interference between individual lights using an optical filter;
an FPGA module that receives image data from the image sensor of the camera device and stores the data classified into 2D and 3D areas in a separate memory space to separately classify 2D image and 3D shape data and transmit the data to the collecting device;
A 2D lighting device consisting of a line-type lighting as a lighting device in the visible light region and consisting of an LED or a laser; and
Concrete crack detection system using 2D and 3D configuration, characterized in that it includes; a 3D laser device composed of a line-type laser as an infrared region illumination device. According to claim 1,
The collection device is
an interface device that collects high-speed image data received from the inspection device, generates one image frame, and transmits the generated image frame data to the data collection and analysis device;
a data collection and analysis device that analyzes and verifies cracks by filtering 2D crack inspection results using 2D analysis for performing crack analysis on the concrete ballast in 2D images and depth information for crack locations in 3D shape data; and
Concrete crack detection system using 2D and 3D configuration, comprising: a distance synchronization device that analyzes the moving distance information of the vehicle received from the tachometer and generates a shooting trigger signal at equal intervals to the inspection device. According to claim 1,
The data collection and analysis device
a data acquisition module that receives an original image of a concrete image taken by an image sensor of a camera device as an image input, and classifies 2D and 3D data into 2D images and 3D shape images, respectively;
a noise removal module that separates the structures and rail tracks of the concrete ballast, which are linear patterns in 2D image and 3D shape data, from the ballast because cracks on the concrete ballast are created on the top of the ballast for ballast recognition;
a filtering module that performs a process for removing a concrete structure and a rail track from a map image using a straight line detection algorithm;
In the crack analysis of 2D images, the standard pattern of cracks is input in advance, and the process of recognizing cracks that match the standard pattern of cracks in the acquired concrete ballast image, and depth information analysis data deeper than the preset depth based on the height information of the upper part of the ballast a crack analysis module that performs the process of extracting ; and
2D comprising a; crack analysis verification module that verifies the analysis of the crack location using height information in 3D inspection to confirm erroneous detection by foreign substances in the 2D image crack analysis process of the crack analysis module and a concrete crack detection system using a 3D configuration. According to claim 1,
The inspection device classifies a region of interest in which only a portion of the image sensor area is acquired into 2D and 3D regions, and images are acquired at equal intervals according to the moving distance of the vehicle, and the tachometer installed on the wheel moves the vehicle. A concrete crack detection system using a 2D and 3D configuration, characterized in that the distance is measured and the location information is synchronized between the 2D and 3D data using an equal interval signal.

Images