A road crack detection and extraction device and detection method thereof

Abstract

The present invention discloses a road crack detection and extraction device and a detection method thereof, including a bracket, a radar antenna, a pan / tilt, a high-speed camera, a sunshade, a light strip and a connecting flange. The beneficial effect of the present invention is that the crack area ratio in the available road damage index can be detected. The device effectively combines the advantages of road radar detection and road image recognition, more accurately identifies and marks the crack position, and improves the image segmentation and image denoising methods according to the crack characteristics, and applies morphology to accelerate the calculation method of the damaged area of ​​the road crack type, effectively improving the crack detection and data processing speed. And the method listed in this application can be extended to similar test detections such as concrete structure crack recognition.

Description

Technical Field

[0001] The invention relates to a road surface crack detection device, in particular to a road surface crack detection and extraction device and a detection method thereof, belonging to the technical field of detection and testing. Background Art

[0002] The daily maintenance and special maintenance of highway projects often require the evaluation of the use status of highways. The Highway Technical Condition Assessment Standard (JTG 5210-2018) stipulates the parameters that need to be tested and assessed. The pavement technical condition PQI is calculated by the pavement damage index PCI, while the PCI of asphalt pavement is calculated by the cracks, rutting, subsidence, wave hugging, potholes, looseness, oil spills, and repairs detected by the pavement survey.

[0003] The most common damage to asphalt pavement is various cracks, and parameters such as crack length need to be measured. The crack length multiplied by 0.2 meters is used as an indicator of the pavement crack disease area. Crack detection investigation is one of the most arduous tasks in road condition investigation. In addition to manual investigation, some automatic detection equipment has also been developed. It mainly uses a vehicle-mounted high-speed camera to take pictures continuously, and then uses computer image recognition to assist manual identification to confirm the cracks and calculate the number of cracks. However, the use of a high-speed camera alone is affected by light, shadows, road debris, and the asphalt pavement's own texture. The misjudgment rate of image recognition is high, and the detection efficiency is still low. Based on this, the present application proposes a pavement crack detection and extraction device and a detection method thereof. Summary of the invention

[0004] The purpose of the present invention is to provide a road crack detection and extraction device and a detection method thereof in order to solve the problem.

[0005] The present invention achieves the above-mentioned purpose through the following technical scheme: a road crack detection and extraction device, including a bracket, a radar antenna, a pan-tilt, a high-speed camera, a light shield, a light strip and a connecting flange; a radar antenna and a pan-tilt are installed at one end of the bracket, a high-speed camera is installed on the pan-tilt, a light shield is fixedly arranged on the cantilever of the bracket, and a light strip is fixedly installed on the inner edge of the light shield, the other end of the bracket is connected to the detection vehicle through a set connecting flange, and the control lines of the pan-tilt and high-speed camera installed on the bracket are all laid along the bracket and connected to the radar data collector and computer in the vehicle, and the computer controls the data collection.

[0006] As a further solution of the present invention: the sunshade is a long strip cover with a length of 3.5 meters and a width of 0.8 meters, and the sunshade uses steel wire strips as a skeleton, the outer layer is a sunshade cloth, and the inner layer is a white reflective cloth.

[0007] As a further solution of the present invention: the light strip adopts a monochromatic LED light, and the light strip is installed in a special light-hiding structure on the inner edge of the sunshade.

[0008] As a further solution of the present invention: the radar antenna is an air-coupled antenna using 1 GHz.

[0009] A detection method for a pavement crack detection and extraction device, the detection method comprising the following steps:

[0010] Step 1: Install the device bracket at the rear of the vehicle, connect the power supply and control line; adjust the pan / tilt and camera shooting angles, and set the camera shooting frequency; the vehicle drives at a constant speed in the detection lane, and the road radar and high-speed camera collect data synchronously;

[0011] Step 2: After data collection is completed, background processing is performed. Based on the detection time synchronization, manual fine-tuning is performed to match the radar reflection wave data with the mileage position represented by the photo image;

[0012] Step 3: The photo image is automatically corrected by software to remove distortion; according to the channel weights specially set based on the color of the LED light, the color photo is converted into a grayscale image according to the weights;

[0013] Step 4: Use smoothing filter and Gaussian curvature filter to remove noise and enhance the edge structure of the image;

[0014] Step 5: The position where the radar reflected wave suddenly exceeds the threshold indicates that there is a transverse crack at that position, and the image narrow strip sample corresponding to the position is extracted, and the gray value distribution map is calculated separately with the background sample;

[0015] Step 6: Calculate the segmentation threshold based on the significant difference in grayscale distribution between the crack sample and the background sample, convert the grayscale image into a binary image and invert it;

[0016] Step 7: Use the "fence connectivity algorithm" to eliminate small blocks such as road texture and spots that interfere with the image;

[0017] Step 8: Based on the feature that cracks are thin and long linear images, morphology is used to analyze the image. Morphological corrosion is performed on the image to restore other corroded images.

[0018] Step 9: Perform morphological dilation on the image, and multiply the counted number of pixels by the area corresponding to each pixel to obtain the area of ​​the crack.

[0019] As a further solution of the present invention: in the step one, the road surface radar and the high-speed camera perform synchronous detection on the same position of the road surface, and the road surface radar uses the crack position in the image with obvious reflection characteristic peaks of the crack to improve the accuracy of image segmentation and recognition of crack image recognition.

[0020] As a further solution of the present invention:.

[0021] The beneficial effects of the present invention are as follows: the pavement crack detection and extraction device and its detection method are reasonably designed, and can be detected by using the crack area ratio in the pavement damage index. The device effectively combines the advantages of pavement radar detection and pavement image recognition, more accurately identifies and marks the crack location, and improves the image segmentation and image denoising methods based on crack characteristics, and applies morphology to accelerate the calculation method of the damaged area of ​​pavement crack types, effectively improving the crack detection and data processing speed. The method listed in this application can be extended to similar test detections such as concrete structure crack recognition. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 It is a schematic diagram of the structure of the detection device of the present invention;

[0023] Figure 2 The figure is a diagram illustrating the crack identification algorithm of the present invention;

[0024] Figure 3 This is a flow chart of the image crack extraction calculation method of the present invention.

[0025] In the figure: 1. bracket, 2. radar antenna, 3. gimbal, 4. high-speed camera, 5. sunshade, 6. light strip and 7. connecting flange. DETAILED DESCRIPTION

[0026] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0027] Embodiment 1

[0028] See also Figure 1 A road crack detection and extraction device comprises a bracket 1, a radar antenna 2, a pan-tilt platform 3, a high-speed camera 4, a light shield 5, a light strip 6 and a connecting flange 7; the radar antenna 2 and the pan-tilt platform 3 are installed at one end of the bracket 1, the high-speed camera 4 is installed on the pan-tilt platform 3, a light shield 5 is fixedly arranged on the cantilever of the bracket 1, and the light strip 6 is fixedly arranged on the inner edge of the light shield 5, the other end of the bracket 1 is connected to the detection vehicle through the arranged connecting flange 7, and the control lines of the pan-tilt platform 3 and the high-speed camera 4 installed on the bracket 1 are all laid along the bracket 1 and connected to the radar data collector and the computer in the vehicle, and the data is collected by the computer control.

[0029] Furthermore, in an embodiment of the present invention, the sunshade 5 is a long strip of cover with a length of 3.5 meters and a width of 0.8 meters, and the sunshade 5 uses steel wire strips as a skeleton to keep it light while requiring a certain rigidity; the outer layer is a blackout cloth and the inner layer is a white reflective cloth.

[0030] Furthermore, in an embodiment of the present invention, the light strip 6 uses a monochrome LED light, and the light strip 6 is installed in a special light-hiding structure on the inner edge of the sunshade 5. The LED light is irradiated on the reflective cloth of the sunshade 5 to form diffuse reflected light and then irradiates the road surface to avoid shadows.

[0031] Furthermore, in an embodiment of the present invention, the radar antenna 2 is an air-coupled antenna of 1 GHz, which is suitable for measuring cracks with a depth of 0 to 20 cm.

[0032] Embodiment 2

[0033] See also Figures 2-3 , a detection method for a pavement crack detection and extraction device, the detection method comprising the following steps:

[0034] Step 1: Install the device bracket at the rear of the vehicle, connect the power supply and control line; adjust the pan / tilt and camera shooting angles, and set the camera shooting frequency; the vehicle drives at a constant speed in the detection lane, and the road radar and high-speed camera collect data synchronously.

[0035] Step 2: After data collection is completed, background processing is performed. Based on the detection time synchronization, manual fine-tuning is performed to match the radar reflection wave data with the mileage position represented by the photo image.

[0036] Step 3: The photo image is automatically corrected by software to remove distortion; according to the channel weights specially set based on the color of the LED light, the color photo is converted into a grayscale image according to the weights.

[0037] Step 4: Use smoothing filtering and Gaussian curvature filtering to remove noise and enhance the image edge structure.

[0038] Step 5: The position where the radar reflected wave suddenly changes beyond the threshold indicates that there is a transverse crack at that location. The narrow strip sample of the image corresponding to the position is extracted, and the gray value distribution map is statistically analyzed separately with the background sample.

[0039] Step 6: Calculate the segmentation threshold based on the significant difference in the grayscale distribution between the crack sample and the background sample, convert the grayscale image into a binary image and invert it, that is, the original darker position of the crack corresponds to 1, and the original brighter position corresponds to 0.

[0040] Step 7: Use the "fence connectivity algorithm" to eliminate small blocks such as road texture and spots that interfere with the image. Take a 7×7 fence as an example. The interior is a 3×3 block. Each pixel on the edge of the block is iThe fence connected to it has at least one pixel b j is also 1, that is, this cluster is connected to the outside of this fence. This algorithm is defined as 8-connected, a2 corresponds to b2, b3, b4; a5 corresponds to b5, b6, b7, b8, b9. If the cluster is not connected to the outside of the fence, the cluster is an isolated "spot" that can be eliminated. The size of the "fence" can be selected for calculation based on the road surface scale represented by each pixel in the image and the scale of the isolated spots that need to be excluded. According to the following formula:

[0041] with a i Adjacent b j

[0042]

[0043] When K is 1, the area inside the fence is an isolated and removable "spot".

[0044] Step 8: Based on the feature that cracks are thin and long linear images, morphology is used for image analysis. The image is morphologically eroded, and the number of erosions is twice the pixel value corresponding to the maximum width of the crack. If there are still pixels with a value of 1 after corrosion, there is an interference image of a large structure at this location. The interference image at this location is removed and other eroded images are restored.

[0045] Step 9: Perform morphological dilation on the image, with the dilation times being the pixel value corresponding to 0.1 meter. The damaged area of ​​the road crack is calculated by multiplying the crack length by 0.2 meters. Therefore, at this time, we only need to count the number of pixels with an image value of 1 and multiply it by the area corresponding to each pixel to get the area of ​​the crack.

[0046] Furthermore, in an embodiment of the present invention, in step one, the road surface radar and the high-speed camera perform synchronous detection on the same position of the road surface, and the road surface radar uses the crack position in the image with obvious reflection characteristic peaks of the crack to improve the accuracy of image segmentation and recognition of crack image recognition.

[0047] Working principle: LED light strips 6 are arranged around the sunshade 5 to illuminate the inner surface of the sunshade 5, forming diffuse reflected light to illuminate the road surface to be inspected; the high-speed camera 4 and the air-coupled antenna of the road surface detection radar are integrated and installed in the middle position of the sunshade 5 to fix the relative detection position.

[0048] Radar road surface detection is more sensitive to transverse cracks in the road surface, and obvious reflection peaks can appear, but the length of the cracks cannot be directly detected. By integrating the radar antenna with a high-speed camera, the area with cracks can be marked on the captured image by comparing the detection time point. At this time, the specially compiled image recognition software can more accurately identify cracks by comparing the grayscale values ​​of the adjacent crack-marked area and the non-crack-marked area.

[0049] It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above and that the invention can be implemented in other specific forms without departing from the spirit or essential features of the invention. Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description, and it is intended that all variations falling within the meaning and scope of the equivalent elements of the claims be included in the invention. Any reference numeral in a claim should not be considered as limiting the claim to which it relates.

[0050] In addition, it should be understood that although the present specification is described according to implementation modes, not every implementation mode contains only one independent technical solution. This description of the specification is only for the sake of clarity. Those skilled in the art should regard the specification as a whole. The technical solutions in each embodiment may also be appropriately combined to form other implementation modes that can be understood by those skilled in the art.

Claims

1. A detection method for a road surface crack detection and extraction device, characterized in that: The pavement crack detection and extraction device comprises a bracket (1), a radar antenna (2), a pan-tilt platform (3), a high-speed camera (4), a light shield (5), a light strip (6) and a connecting flange (7); the characteristics are as follows: the radar antenna (2) and the pan-tilt platform (3) are installed at one end of the bracket (1), the high-speed camera (4) is installed on the pan-tilt platform (3), a light shield (5) is fixedly arranged on the cantilever of the bracket (1), and the light strip (6) is fixedly arranged on the inner edge of the light shield (5), the other end of the bracket (1) is connected to the detection vehicle through the provided connecting flange (7), and the control lines of the pan-tilt platform (3) and the high-speed camera (4) installed on the bracket (1) are all laid along the bracket (1) and connected to the radar data collector and the computer in the vehicle, and the computer controls the data collection; The sunshade (5) is a long strip-shaped cover with a length of 3.5 meters and a width of 0.8 meters, and the sunshade (5) uses steel wire strips as a frame, an outer layer is a light-shielding cloth, and an inner layer is a white reflective cloth; the light strip (6) uses a single-color LED light, and the light strip (6) is installed in a special light-hiding structure at the inner edge of the sunshade (5); the radar antenna (2) uses a 1 GHz air-coupled antenna; The detection method comprises the following steps: Step 1: Install the device bracket at the rear of the vehicle, connect the power supply and control line; adjust the pan / tilt and camera shooting angles, and set the camera shooting frequency; the vehicle drives at a constant speed in the detection lane, and the road radar and high-speed camera collect data synchronously; Step 2: After data collection is completed, background processing is performed. Based on the detection time synchronization, manual fine-tuning is performed to match the radar reflection wave data with the mileage position represented by the photo image; Step 3: The photo image is automatically corrected by software to remove distortion; according to the channel weights specially set based on the color of the LED light, the color photo is converted into a grayscale image according to the weights; Step 4: Use smoothing filter and Gaussian curvature filter to remove noise and enhance the edge structure of the image; Step 5: The position where the radar reflected wave suddenly exceeds the threshold indicates that there is a transverse crack at that position, and the image narrow strip sample corresponding to the position is extracted, and the gray value distribution map is calculated separately with the background sample; Step 6: Calculate the segmentation threshold based on the significant difference in grayscale distribution between the crack sample and the background sample, convert the grayscale image into a binary image and invert it; Step 7: Use the "fence connectivity algorithm" to eliminate small blocks such as road texture and spots that interfere with the image. Step 8: Based on the feature that the crack is a thin and long linear image, morphology is used to analyze the image, and morphological corrosion processing is performed on the image to restore other corroded images; Step 9: Perform morphological dilation on the image, and multiply the counted number of pixels by the area corresponding to each pixel to obtain the area of ​​the crack; In the step 1, the road surface radar and the high-speed camera perform synchronous detection on the same position of the road surface, and the road surface radar uses the cracks to mark the crack positions with obvious reflection characteristic peaks.

Images