An image registration method for visualizing bridge inspection

Abstract

The application belongs to the technical field of bridge engineering test detection, and relates to an image registration method for visual bridge detection, which is characterized by comprising the following steps: collecting device selection, collecting device parameter extraction, calculating collecting overlap rate, GPS control point layout, total station calibration point coordinate collection, beam body image collection, giving image calibration point coordinates, and image registration. Compared with the conventional post-calibration method, the image registration method for visual bridge detection is more accurate, and absolute coordinate system is given, so that the whole bridge can form a whole, and the position, length, width and other information of the disease can be accurately obtained.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of bridge engineering test detection, and relates to an image splicing method for bridge body appearance detection, in particular to an image registration method for visual bridge detection. BACKGROUND

[0002] In the field of bridge engineering test detection, due to the special structure of the beam body, the GPS positioning accuracy under the bridge is very poor, it is difficult to provide high-precision positioning information for the image acquisition device, and only one image in multiple images can be used as a reference image, the coordinate system of the other related images is arbitrary, and the unit of the image splicing is pixel, which causes many inconveniences in subsequent disease identification, labeling and statistical work. The original film with sufficient overlap rate and high resolution, and the high-precision positioning information of the image required for image registration are the key elements of image registration, and how to collect and register according to the established scheme requirements is a problem that must be solved in visual bridge detection.

[0003] At present, the main method of image registration is feature method, similarity measurement matching is performed on the extracted feature points, then the spatial coordinates of the images are positioned through the matched feature points, the transformation parameters of the spatial coordinates are calculated, and finally the image registration is performed by the coordinate transformation parameters. Due to the characteristics of the above method and the problem that the color of the beam body surface is relatively single, the matching quality, accuracy and reliability are difficult to control. SUMMARY

[0004] The purpose of the present application is to provide an image registration method for visual bridge detection, which can solve the problems of image acquisition quality before image registration, assignment of target coordinate system and conversion of units of meters after image splicing.

[0005] The technical problem of the present application is solved by the following technical scheme:

[0006] An image registration method for visual bridge detection, characterized by comprising the following steps:

[0007] Step S1: selecting the acquisition device according to the required acquisition resolution and the clearance of the beam body;

[0008] Step S2: extracting the parameters of the acquisition device;

[0009] Step S3: obtaining the basic information of the bridge, determining the target resolution and overlap rate of the acquisition, calculating the number of shots in the longitudinal and transverse directions of the bridge, and obtaining the shooting interval in the longitudinal and transverse directions; calculating the distance from the camera to the object from the target resolution, and selecting the lens in the acquisition device in step S1;

[0010] Step S4, using GPS to establish a map root control point on the outside of the bridge, for the station measurement of the total station;

[0011] Step S5, collecting the coordinates of the total station calibration points;

[0012] Step S6, collecting the beam image;

[0013] Step S7, marking the calibration points in the beam image photos, ensuring that the marked photos are not less than 4, and giving the corresponding three-dimensional coordinate information of the point number;

[0014] Step S8, registering the images, calculating the image external parameters and coordinate transformation parameters from the calibration point coordinates, and finally performing spatial triangle calculation to obtain the beam overall image results of the absolute coordinate system.

[0015] Moreover, the collection device comprises a single-lens reflex camera, a lens, and a holder.

[0016] Moreover, the method for selecting the collection device according to the required collection resolution and the clearance of the beam comprises the following steps: determining the single-lens reflex camera pixel to be not less than 30 million pixels in combination with tests and pixel point spacings; selecting a short-focus or fixed-focus lens when the clearance is less than or equal to 6.0 m; and selecting a long-focus lens when the clearance is greater than 6.0 m.

[0017] The holder is selected to be a holder device for shooting vertically upward by 90°.

[0018] Moreover, the parameters of the single-lens reflex camera and the lens comprise a sensor size, effective pixels, a focal length, and an image size.

[0019] Moreover, the bridge basic information comprises a beam length, a beam width, and a clearance.

[0020] Moreover, the target resolution is determined according to the limit value of the crack component and the 0.15 mm demarcation value of the crack repair scheme; and the overlap rate is between 40% and 50%.

[0021] Moreover, the method for collecting the total station calibration point coordinates comprises the following steps: pasting calibration points on a pier column or a bridge structure, marking the numbers on each calibration point, giving the corresponding coordinate values of the calibration points in the photos, and erecting the total station at a control point or under the bridge, using the rear view or the rear resection orientation, collecting all the three-dimensional coordinates of the calibration points through multiple stationing.

[0022] Moreover, the method for collecting the beam image comprises a beam shooting method and a calibration point shooting method, and the beam shooting method comprises the following steps:

[0023] Using the longitudinal and transverse shooting interval calculated in step 3, the grid position of the shooting point under the bridge is determined, and is marked; after adjusting the equipment and setting the camera focal length and the angle of 90° vertically upward, the image acquisition work of the beam body is carried out along the "arch" type shooting route of first longitudinal and then transverse;

[0024] The calibration point shooting method comprises the following steps: using the same equipment to shoot at a wide angle covering the beam bottom and multiple calibration points, ensuring that the calibration points have multiple photos of different angles connected with the beam bottom.

[0025] The advantages and beneficial effects of the present application are:

[0026] The image registration method of the visual bridge detection can improve the quality, precision and reliability of image matching by setting a reasonable optimal overlap rate value for image acquisition; the longitudinal and transverse shooting intervals are calculated by determining the target resolution and lens parameters, thereby meeting the requirements of crack width analysis; the results after image stitching form a whole digital model, meeting the needs of multi-period data comparison; the image registration carries out the calibration point piercing, and realizes the image of the absolute coordinate system formed by stitching. Compared with the conventional post-calibration method, the present application has higher precision and absolute coordinate system, and can form a whole bridge, and can accurately obtain the position, length, width and other information of the disease. BRIEF DESCRIPTION OF DRAWINGS

[0027] Figure 1 The method flowchart of the present application is shown in the figure;

[0028] Figure 2 The pattern of the calibration point of the present application is shown in the figure (style one is a triangular intersection figure, style two is a circular triangular intersection figure, style three is a square triangular intersection figure, and style four is an L-shaped figure). DETAILED DESCRIPTION

[0029] The present application will be further described in detail below through specific embodiments, and the following embodiments are only descriptive and not limiting, and cannot limit the protection scope of the present application.

[0030] The present application provides an image registration method of visual bridge detection, which is characterized by: comprising the steps of selecting a collection device, extracting a collection device parameter, calculating a collection overlap rate, laying out a GPS control point, collecting a calibration point coordinate of a total station, collecting a beam body image, assigning a calibration point coordinate in the image, and image registration.

[0031] Step S1: According to the basic information of the bridge, the resolution of the collection and the clearance of the beam body, select the appropriate single-lens reflex camera, lens and gimbal. Combined with the test and the pixel spacing, it is determined that the pixel of the single-lens reflex camera should not be less than 30 million. When the clearance is ≤6.0m, select a short-focus or fixed-focus lens; when the clearance is >6.0m, select a long-focus lens. The gimbal selects a gimbal device that can shoot vertically upward by 90°. It is recommended to select DJI or Zhiyun, which has the characteristics of strong compatibility and expansion capability;

[0032] Step S2: Extract the basic parameters of the camera and lens, including: sensor size (mm), effective pixels, focal length (mm), image size (pixels);

[0033] Step S3: Obtain the basic information of the bridge, including: beam length, beam width, clearance. Determine the target resolution and overlap rate. The target resolution is determined according to the limit value of the crack component and the 0.15mm demarcation value of the crack repair scheme. The overlap rate is obtained from the test. When it is less than 40%, it is easy to cause fusion error and failure. When it is greater than 50%, it causes low efficiency and large data file. The reasonable overlap rate is between 40% and 50%. From the above information and the target resolution and overlap rate, the number of shots in the longitudinal and transverse directions of the bridge can be calculated, so as to obtain the shooting interval in the longitudinal and transverse directions. From the target resolution, the distance between the camera and the object can be calculated, so as to select the lens in step S1;

[0034] Step S4: Use GPS to establish a graph root control point on the outside of the bridge. Each hole is equipped with not less than one control point for station measurement of the total station. The coordinate system is determined according to the scheme or the local coordinate system to ensure that the coordinate system of the whole bridge is the same;

[0035] Step S5: Paste the calibration points on the pier column or bridge structure, as shown in Figure 2 The specification is 10*10cm square, and the pattern has a triangular cross pattern, a circular triangular cross pattern, a square triangular cross pattern, and an L-shaped pattern. Each calibration point indicates the number on the top to facilitate the assignment of corresponding coordinate values to the calibration points in the photos. The total station is erected at the control point or under the bridge, and uses the rear view or rear intersection orientation. Through multiple stationing, the three-dimensional coordinates of all calibration points are collected;

[0036] Step S6: ① Beam body shooting: using the longitudinal and transverse shooting interval calculated in step 3, the grid position of the shooting point under the bridge is determined and marked. After adjusting the equipment and setting the camera focus and the angle of 90° vertically upward, the image collection work of the beam body is carried out along the "bow" type shooting route of first longitudinal and then transverse. ② Calibration point shooting: using the same equipment, wide-angle shooting is carried out in the whole area, as far as possible to cover the beam bottom and multiple calibration points, to ensure that the calibration points and the beam bottom have multiple different angle shooting photos for connection. The shooting method of "from inside to outside, from near to far" is adopted, the inner layer is close-range crack collection work, and the outer layer is long-distance overall wide-angle collection work;

[0037] Step S7: the calibration points are punctured in the image photos, to ensure that the punctured photos are not less than 4, and to give the corresponding point number three-dimensional coordinate information;

[0038] Step S8: through registration, the feature points of all images are extracted, the matching feature point pairs are found by similarity measurement, the image external parameters and coordinate transformation parameters are calculated from the calibration point coordinates, and finally the spatial triangle calculation is carried out to obtain the beam body overall image results of the absolute coordinate system.

[0039] Although the embodiments of the present application and the drawings are disclosed for the purpose of illustration, those skilled in the art can understand that various substitutions, changes and modifications are possible without departing from the spirit and scope of the present application and the appended claims, therefore, the scope of the present application is not limited to the disclosed content of the embodiments and the drawings.

Claims

1. An image registration method for visual bridge detection, characterized in that: Includes the following steps: Step S1: Select the acquisition device based on the required acquisition resolution and the clearance of the beam. Step S2: Extract the parameters from the acquisition device; Step S3: Obtain basic bridge information, determine the target resolution and overlap rate, and calculate the number of longitudinal and transverse shots of the bridge to obtain the longitudinal and transverse shooting intervals. The distance between the camera and the object being photographed is calculated from the target resolution, thereby selecting the lens in the acquisition device in step S1; Step S4: Use GPS to establish control points on the outside of the bridge for total station surveying. Step S5: Collect the coordinates of the total station calibration points; Step S6: Acquire images of the beam. Step S7: In the photos of the beam image, the calibration points are marked with pins, ensuring that there are no fewer than 4 photos of the pins, and assigning the three-dimensional coordinate information of the corresponding point number. Step S8 involves registering the image, calculating the image's extrinsic parameters and coordinate transformation parameters from the calibration point coordinates, and finally performing spatial triangulation to obtain the overall image of the beam in the absolute coordinate system.

2. The image registration method for visual bridge detection according to claim 1, characterized in that: The acquisition equipment includes a DSLR camera, a lens, and a gimbal.

3. The image registration method for visual bridge detection according to claim 2, characterized in that: The method for selecting the acquisition equipment based on the required acquisition resolution and the beam clearance is as follows: The selection is determined by combining experiments and pixel spacing; the SLR camera should have at least 30 megapixels; when the clearance is ≤6.0m, a short-focus or fixed-focus lens is selected; when the clearance is >6.0m, a telephoto lens is selected. Choose a gimbal that shoots vertically upwards at a 90° angle.

4. The image registration method for visual bridge detection according to claim 2, characterized in that: The parameters extracted from SLR cameras and lenses include: sensor size, effective pixels, focal length, and image size.

5. The image registration method for visual bridge detection according to claim 1, characterized in that: The basic information of the bridge includes: beam length, beam width, and clearance.

6. The image registration method for visual bridge detection according to claim 1, characterized in that: The target resolution is determined based on the limit value of the cracked component and the 0.15mm dividing value of the crack repair scheme; the overlap rate is between 40% and 50%.

7. The image registration method for visual bridge detection according to claim 1, characterized in that: The method for collecting the coordinates of the total station calibration points is as follows: calibration points are pasted on the piers or bridge structures, and each calibration point is marked with a number. The calibration points in the photograph are assigned corresponding coordinate values. The total station is set up at the control point or under the bridge, and backsight or resection orientation is used. Through multiple station setups, the three-dimensional coordinates of all calibration points are collected.

8. The image registration method for visual bridge detection according to claim 1, characterized in that: The method for acquiring beam images includes a beam imaging method and a calibration point imaging method. The beam imaging method includes the following steps: Using the longitudinal and transverse shooting intervals calculated in step 3, determine the grid positions of the shooting points under the bridge and mark them; after debugging the equipment and setting the camera focal length and vertical upward 90° angle, carry out beam image acquisition work along the "bow" shaped shooting route first longitudinally and then laterally. The calibration point shooting method includes the following steps: using the same equipment to take wide-angle shots of the entire area, covering the bottom of the beam and multiple calibration points, and ensuring that multiple photos taken from different angles are connected between the calibration points and the bottom of the beam.

Images