CN107941147A - Large scale system three-dimensional coordinate non-contact forecasting method - Google Patents

Abstract

The invention provides a method for non-contact on-line measurement of three-dimensional coordinates of a large-scale system. The measuring device used includes a camera, a laser displacement sensor, a projector and a control system; The system is connected, and the laser displacement sensor is used to scan the dead angle of the measured object with high precision, and the laser displacement sensor transmits the scanned image to the control system. The invention adopts the fusion of monocular vision information, structured light and laser displacement detection data, can measure three-dimensional dimensions with high precision, and can be used for state monitoring and quality inspection during processing of key components with large complex curved surfaces and complex structures.

Description

Non-contact online measurement method of three-dimensional coordinates for large-scale systems

technical field

The invention relates to the field of three-dimensional dimension measurement, in particular to a non-contact online measurement method for three-dimensional coordinates of a large system.

Background technique

In military and national defense fields such as aerospace, missiles, and heavy weapons and equipment, with the increasingly high requirements for the performance of large-scale systems, the processing and manufacturing of key components with large complex surfaces and complex structures, and the 3D measurement of monitoring the state and performance of large components have received great attention. Unprecedented challenges and opportunities. In the fields of automobile and ship manufacturing, China's production and sales have been at the forefront of the world in recent years. However, corresponding to the strong market demand and production and sales, the domestic large-scale ship propellers and manufacturing capacity are seriously insufficient, and the independent shipment rate is low; Precision molds and high-performance engines mainly rely on imports, which seriously restricts the independent development of my country's equipment manufacturing and basic industries. The root cause of these problems is that the three-dimensional dimension measurement level of key components with large complex curved surfaces and complex structures is relatively backward, resulting in low manufacturing efficiency and difficulty in guaranteeing machining accuracy.

As an important support for modern manufacturing, measurement technology is not only used as a means of evaluating the quality of final products, but more importantly, it serves for the quality monitoring of product design and manufacturing processes, and provides complete information feedback for the manufacturing process. The measurement method has also changed from the traditional "off-line" measurement to the manufacturing process to realize "on-line" measurement. Compared with offline measurement, online measurement has great advantages: online measurement measures the workpiece at the processing site, no need to manufacture additional measurement fixtures and transfer objects to the measurement room, which can reduce economic and time consumption; online measurement moves the measurement process to In the process of processing, the quality control is integrated into the manufacturing process, and the quality inspection after processing is changed into real-time inspection during processing. Quality problems can be found at any time and the process can be adjusted in time to improve the yield of products; online measurement can detect defects in the manufacturing process in time. Problems, and correct these problems in time by adjusting the processing parameters and introducing error compensation, so as to realize agile processing and ensure processing quality. Therefore, how to measure the three-dimensional dimensions of key components with large complex surfaces and complex structures online has become an important issue in the fields of military and equipment manufacturing such as aerospace, missiles, weaponry, automobiles, and shipbuilding.

Contents of the invention

The purpose of the present invention is to provide a method for non-contact online measurement of three-dimensional coordinates of a large-scale system to solve the problem of large errors in online measurement of three-dimensional dimensions of components with large complex curved surfaces and complex structures.

Specifically, the present invention provides a non-contact online measurement method for three-dimensional coordinates of a large system, which includes using a laser displacement sensor to perform high-precision scanning of the dead corner of the measured object, and the laser displacement sensor transmits the scanned image to the control system.

Since the laser displacement sensor has high measurement accuracy, using it to measure the shape can achieve high-precision measurement with low cost.

Further, the method includes the following steps:

Step 1. Calibrate the camera;

Step 2, the projector projects the periodically modulated fringe pattern onto the object to be measured;

Step 3, the camera shoots the fringe pattern on the measured object;

Step 4, the control system processes the image captured by the camera;

Step five, the laser displacement sensor performs high-precision scanning;

Step 6, the control system performs three-dimensional image reconstruction.

The method can realize non-contact online measurement of large-scale systems, has the characteristics of high measurement accuracy, and can be used for state monitoring and quality inspection of key components with large complex surfaces and complex structures during processing.

Further, the devices used in the measurement process include cameras, laser displacement sensors, projectors and control systems;

The camera, laser displacement sensor, and projector are respectively connected to the control system through cables;

According to the control instructions of the control system, the projector projects the sent periodic modulated fringe pattern onto the surface of the large-scale system to be measured;

The camera shoots the fringe patterns on multiple parts of the measured object, and transmits the captured images to the control system;

The laser displacement sensor performs high-precision scanning at the dead angle of the measured object, and transmits the scanned image to the control system;

The control system first processes the fringe pattern and the captured image to obtain the basic three-dimensional topography of the system to be tested; combined with the image scanned by the laser displacement sensor, the final three-dimensional size and image of the system to be tested are obtained.

The projector projects stripes, and the camera and laser displacement sensor cooperate to take pictures of the measured object to obtain high-precision and clear pictures, which are further processed by the control system.

Further, step 1 includes using the calibration method of the 2D plane template to calibrate the camera parameters; using the standard plane translation method to calibrate the phase-height of the camera.

Camera calibration is used to improve the accuracy of the camera and improve the clarity and accuracy of imaging.

Further, the camera parameters include focal length f, pixel spacing Δx in the horizontal direction of the image plane, pixel spacing Δy in the vertical direction of the image plane, and principal point coordinates (u ₀ , v ₀ ).

The calibration of camera parameters plays an important role in camera performance.

Further, the phase-height mapping relationship is: in is the phase, h is the height, and a and b are constants.

After the calibration of the phase-height mapping relationship is completed, the camera imaging accuracy is high, which can improve the accuracy of the system's online measurement.

Further, the periodically modulated fringe pattern is a sinusoidal fringe pattern.

Periodically modulated fringes are stable and have certain regularity. When they are projected onto an object and change, the actual pattern of the object can be obtained according to the deformation and existing imaging rules.

Further, Step 4 includes that the control system uses a four-step phase shift method to extract the phase and uses a phase unwrapping algorithm to unwrap the phase.

The control system fuses the graphics to obtain the three-dimensional data of the object.

The beneficial effects of the present invention are:

(1) Using monocular visual information, structured light and laser displacement detection data fusion, it can measure three-dimensional dimensions with high precision;

(2) It can be used for status monitoring and quality inspection of key components with large complex surfaces and complex structures;

(3) The laser displacement sensor is used for scanning, which has high measurement accuracy, and it is used to measure the shape, which can realize high-precision measurement and low cost.

Description of drawings

The drawings are for the purpose of illustrating specific embodiments only and are not to be considered as limitations of the invention, and like reference numerals refer to like parts throughout the drawings.

Fig. 1 is a schematic diagram of the overall structure of the online measuring device of the present invention;

Fig. 2 is the flow chart of on-line measurement method of the present invention;

Fig. 3 is the image captured by camera in the embodiment of the present invention;

Fig. 4 is the graph that obtains after image unwrapping processing in the embodiment of the present invention;

Fig. 5 is a graph obtained by three-dimensional reconstruction in an embodiment of the present invention.

In the figure: 1-projector, 2-camera, 3-control system, 4-laser displacement sensor.

Detailed ways

The present invention will be further described in detail below in conjunction with the accompanying drawings and specific embodiments.

For large ^- scale systems with an area greater than 25m2, the present invention provides a three-dimensional coordinate non-contact online measurement method, as shown in Figure 1, the devices used in the implementation of the method include cameras, laser displacement sensors, projectors and control system, etc.

The camera, the laser displacement sensor and the projector are respectively connected with the control system through cables.

According to the control instructions of the control system, the projector projects the sent periodic modulated fringe pattern onto the surface of the large-scale system to be measured;

The camera shoots the fringe patterns on multiple parts of the measured object, and transmits the captured images to the control system;

The laser displacement sensor performs high-precision scanning at the dead angle of the measured object, and transmits the scanned image to the control system;

The control system first processes the fringe pattern and the captured image to obtain the basic three-dimensional topography of the system to be tested; combined with the image scanned by the laser displacement sensor, the final three-dimensional size and image of the system to be tested are obtained.

Specifically, the large-scale system three-dimensional coordinate non-contact online measurement method, as shown in Figure 2, includes the following steps:

Step 1. Calibrate the camera

Install the camera and projector, adjust the relative position and optical axis angle between the camera and projector, and ensure that the device under test is within the projection range of the projector. The calibration method of the 2D plane template is used to calibrate the camera parameters; the phase-height mapping relationship of the camera is obtained by using the standard plane translation method, and the phase-height calibration is realized.

In order to ensure that the device under test is within the projection range of the projector, the number of projectors installed is n, and all projectors are installed on a spherical surface with the center of the device as a spherical shape and a radius of R, and the distribution is uniform. When n( When R×α)≥2S, it can be considered that the device under test is within the projection range of the projector, where S is the surface area of the device, and α is the projection angle of the projector.

The camera parameters include the focal length f, the pixel spacing Δx in the horizontal direction of the image plane, the pixel spacing Δy in the vertical direction of the image plane, and the coordinates of the projection position of the camera optical center on the CCD imaging plane, that is, the principal point coordinates (u ₀ , v ₀ ). Because the relationship between the three-dimensional geometric position of a point on the surface of the object to be measured and its corresponding point in the image coordinate system is determined by the position of the camera in space and the parameters of the camera, it is necessary to accurately calculate each point of the object The 3D geometric position of , the camera parameters must be calibrated first.

The phase-height mapping relationship is: in is the phase, h is the height, and a and b are constants.

Step 2. Projector Projection

In the traditional phase shift method, the measurement system needs a relatively complex phase shift device to control and generate precise fixed-step phase shift, which is relatively complicated. In the method, the control system compiles and generates the periodic modulation fringe pattern for projection, which is projected onto the object to be measured by a projector. The control system issues a projection command, and uses a single projector to project the periodically modulated fringe pattern onto the part of the surface of the measured object. Due to the change of the height of the part of the object, the phase of the grating fringes at each point is shifted.

Step 3: Camera to shoot

The camera quickly shoots the fringe images on multiple parts of the measured object to obtain multiple fringe images with phase difference, and the images contain the surface information of the measured part. The camera sends the captured picture information to the control system.

Step 4. The control system processes the images captured by the camera

After all the projectors are projected and photographed once, according to the captured fringe image containing the surface information of the measured part, the control system uses the four-step phase shift method to extract the phase, and then uses the phase unwrapping algorithm to unwrap the phase.

According to the pre-calibrated camera parameters and phase-height mapping relationship in step 1, after the phase is obtained during actual measurement, according to the determined phase-height relationship, the three-dimensional height point data of the object surface is recovered from the phase value, according to the phase-height relationship Height mapping relational expression, the measured phase can be substituted into the relational expression to calculate the height value. Complete the measurement of the three-dimensional topography of the object surface and the basic three-dimensional reconstruction of the measured object, that is, obtain the three-dimensional images of each curved surface. The control system adopts a non-mark point splicing algorithm to carry out reasonable image splicing of the three-dimensional images of each curved surface, and obtains the basic three-dimensional topography map of the measured object.

Phase unwrapping algorithm: first use the four-step phase shift to get the packed phase, and then use the region growing method to unwrap.

The phase values obtained by using the four-step phase shift method are wrapped in the value range [-π, π] of the arctangent function, and there are discontinuous phases with discontinuities and jumps. The phase does not directly reflect the real phase change, only the phase unwrapping can eliminate the discontinuities and jumps in the phase principal value diagram, and obtain the continuous distribution of the true phase value. Unwrapping process: Use the region growing method to start from the starting point of the phase value, and compare the phase principal values of two adjacent pixels along the row or column of the phase matrix. For a continuously changing object, its phase value should also be Continuous, so when the phase difference between two adjacent pixels exceeds a certain range, add (or subtract) the phase value of an integer multiple of 2π to the phase value of the next pixel to obtain the real phase.

Markerless point stitching algorithm: Due to the limitation of the projected phase field and the shooting range of the camera, it is necessary to perform point cloud stitching after multiple measurements to complete the measurement. The principle of SIFT registration is used to assist in the automatic stitching of point clouds. All scales and image locations are first searched, and potential feature points can be efficiently detected using the Gaussian difference formula, which are invariant to scale scaling and rotation changes. Determine the position and scale of the key points, and assign a direction to each key point. All operations on image data are converted into operations on the direction, scale, and position of feature points, through the gradient of the area around the current scale of the key point Statistics, generating feature point descriptors.

Step 5. Laser displacement sensor for high-precision scanning

In order to ensure that the measurement accuracy meets the requirements, according to the initially obtained 3D image information of large components and the 3D information of each part of the surface, the control system needs to control the information missing points, feature points, and complex structure positions in the images obtained after preliminary splicing. The individual information in the initially obtained 3D image is supplemented and perfected. Specifically, the control system controls the laser displacement sensor as an optical probe, and performs high-precision scanning on the dead corner of the measured object based on the laser displacement detection technology. The laser displacement sensor will scan the obtained The image and displacement information are transmitted to the control system.

The lack of information points, feature points, and complex structure positions in the image are identified by using the Gaussian Laplacian operator detection method.

Step 6. The control system performs 3D image reconstruction

After the high-precision image and displacement information scanned by the laser displacement sensor are transmitted to the control system, the control system adds the data to the existing database, and then uses phase measurement profilometry (PMP) to measure the three-dimensional dimensions of key components with large complex surfaces and complex structures. ) to reconstruct the three-dimensional image to obtain the three-dimensional coordinates of the object, that is, the length, width and height of the object. Obtain 3D dimensions and images that meet the accuracy requirements, and feed back the information into a form readable by the CAM system, thereby realizing non-contact online measurement of 3D dimensions of key components with large complex surfaces and complex structures.

In this embodiment, the three-dimensional coordinate non-contact online measurement is performed on the spherical mirror system as shown in the figure, specifically:

Step 1. Calibrate the camera

The calibration method of the 2D plane template is used to calibrate the camera parameters, and the camera parameters are obtained: Δx=Δy=4.65 μm, f=35.17mm, (u ₀ , v ₀ )=(695.5, 519.5), and the unit is pixel.

Using the standard plane translation method, a=0.3 and b=0.025 in the phase-height mapping relationship are obtained, namely

Step 2. Projector Projection

The control system issues a projection instruction, and the projector is used to project the sinusoidal fringe image onto a part of the curved surface of the measured object.

Step 3: Use the camera to shoot

Use a CCD camera as an image acquisition tool, as shown in Figure 3, to shoot the fringe images on multiple parts of the measured object to obtain multiple fringe images with phase difference, and the CCD camera will send the captured image information to the control panel. system.

Step 4. The control system processes the images captured by the camera

According to the captured fringe image containing the surface information of the measured part, the control system uses the four-step phase shifting method to extract the phase, and then uses the phase unwrapping algorithm to unwrap the phase. The graph obtained after unwrapping is shown in Figure 4 .

Step 5. Laser displacement sensor for high-precision scanning

For the missing points in the three-dimensional image obtained by preliminary stitching, the control system controls the laser displacement sensor to further scan the measured object, and transmits the scanned image to the control system.

Step 6. The control system performs 3D image reconstruction

On the basis of the initial stitching of the 3D images, the control system reconstructs the 3D images according to the images scanned by the laser displacement sensor. The reconstructed 3D images are shown in Figure 5.

In summary, the present invention provides a non-contact online measurement method for three-dimensional coordinates of a large-scale system, which uses monocular vision information, structured light and laser displacement detection data fusion, can measure three-dimensional dimensions with high precision, and can be used for large complex curved surfaces , Condition monitoring and quality inspection of key components with complex structures during processing.

Although the present invention has been described in detail in conjunction with preferred embodiments, those skilled in the art should understand that various modifications are allowed without departing from the spirit and essence of the present invention, and they all fall into the present invention. within the scope of protection of the claims.

Claims (8)

1. The non-contact online measurement method for three-dimensional coordinates of a large system, characterized in that it includes: The laser displacement sensor is used to scan the dead angle of the measured object with high precision, and the laser displacement sensor transmits the scanned image to the control system. 2. The online measuring method according to claim 1, characterized in that, comprising the following steps: Step 1. Calibrate the camera; Step 2, the projector projects the periodically modulated fringe pattern onto the object to be measured; Step 3, the camera shoots the fringe pattern on the measured object; Step 4, the control system processes the image captured by the camera; Step five, the laser displacement sensor performs high-precision scanning; Step 6, the control system performs three-dimensional image reconstruction. 3. The online measurement method according to claim 1 or 2, wherein the devices used in the measurement process include a camera, a laser displacement sensor, a projector and a control system; The camera, laser displacement sensor, and projector are respectively connected to the control system through cables; According to the control instructions of the control system, the projector projects the sent periodic modulated fringe pattern onto the surface of the large-scale system to be measured; The camera shoots the fringe patterns on multiple parts of the measured object, and transmits the captured images to the control system; The laser displacement sensor performs high-precision scanning at the dead angle of the measured object, and transmits the scanned image to the control system; The control system first processes the fringe pattern and the captured image to obtain the basic three-dimensional topography of the system to be tested; combined with the image scanned by the laser displacement sensor, the final three-dimensional size and image of the system to be tested are obtained. 4. The on-line measurement method according to claim 3, wherein said step 1 comprises the calibration method using a 2D plane template to calibrate the camera parameters; adopting the standard plane translation method to determine the phase-height of the camera calibration. 5. online measurement method according to claim 4, is characterized in that, described camera parameter comprises focal length f, the interpixel spacing Δx of image plane horizontal direction, the interpixel spacing Δy of image plane vertical direction, principal point coordinate ( u ₀ , v ₀ ). 6. The online measurement method according to claim 4 or 5, wherein the phase-height mapping relationship is: in is the phase, h is the height, and a and b are constants. 7. The online measurement method according to claim 4 or 5, characterized in that the periodically modulated fringe pattern is a sinusoidal fringe pattern. 8 . The online measurement method according to claim 7 , wherein the step 4 includes the control system extracting the phase using a four-step phase shift method and unwrapping the phase with a phase unwrapping algorithm. 9 .