CN106441094A - Adaptive calibration vision online detection device and method - Google Patents

Abstract

The invention relates to a visual online detection device and method for self-adaptive calibration. The device components include a base, a support plate, a triangular rib plate, a lead screw platform, a stepping motor, an electric rotary platform, a camera mounting plate and an industrial camera and other components. The device realizes the horizontal movement of the camera along the Y and Z axes by two stepping motors, and realizes the rotation of the camera around three coordinate axes by three electric rotating tables. Through the adjustment functions such as up and down, left and right, and rotation, the pose of the industrial camera can be easily adjusted for camera calibration. After the calibration is completed, it automatically enters the object detection stage, and performs operations such as image acquisition, image preprocessing, image segmentation, and morphological processing. After the image detection and measurement is completed, if the error of the detection value of the object to be measured is not within the error range, the calibration will be restarted. The self-adaptive calibration visual on-line detection device has simple and practical structure, convenient operation, high calibration precision, and is stable and reliable. The calibration process is fully automatic without manual intervention, eliminating the interference and errors caused by manual factors.

Description

Adaptive Calibration Visual Online Inspection Device and Method

technical field

The invention relates to the field of automatic detection, in particular to a self-adaptive calibration visual on-line detection device and method.

Background technique

In recent years, with the proposal of China's Industry 4.0 strategy, industrial automation technology is bound to develop rapidly. Machine vision technology is an important direction of industrial automation technology. Machine vision technology is becoming more and more mature, and it is increasingly playing an important role in social production and life, such as visual monitoring, automatic recognition and measurement of parts, 3D reconstruction, terrain matching, medical image processing, etc. It can be said that the camera is the eye of industrial automation, but calibration is related to the precision and accuracy of the eye. Camera calibration is an essential and important step in most machine vision applications, which directly has an important impact on the accuracy of subsequent work. One of the research hotspots in the field of machine vision today is to study various convenient, practical, flexible and high-precision camera calibration systems. Therefore, it is of great practical value to the automation industry to develop a camera calibration system that is easy to operate, fast and convenient. At present, in terms of visual calibration, it is basically the calibration of the internal parameters of the camera in the offline state. For the visual inspection system, the calibration of the internal parameters of the camera can improve the accuracy of the measurement, but the online measurement also involves external parameters such as the attitude and position of the device. Parameters are calibrated and adjusted to ensure the precise and safe operation of the measurement system and adapt to different applications. At present, the adjustment and calibration of the external parameters and the initial state of the device are all manually operated. In addition to the troublesome operation and the need for skilled workers to operate, the calibration results often cannot avoid errors caused by human factors. Therefore, the present invention proposes a device that can adapt to different application places. The device itself can have a posture adjustment function according to the characteristics of the measurement object, and can automatically realize the calibration and online measurement of the internal and external parameters of the vision system.

Contents of the invention

The purpose of the present invention is to design a visual online detection device and method for self-adaptive calibration in view of the deficiencies in the prior art, so that the camera can automatically perform the pose adjustment and calibration process, effectively solving the complicated and easy manual calibration process at present. Errors and other drawbacks, greatly improving the calibration accuracy. At the same time, the time and manpower consumed by additional manual calibration can be saved.

In order to achieve the above object, the idea of the present invention is to perform camera calibration by fixing the calibration plate and rotating the camera. The installation device of the camera is designed as an automatic adjustment device, so as to ensure the free adjustment of the camera attitude during the calibration process. Its main design functions are as follows:

(1) Camera fixation: When the industrial camera is running in the system, tiny vibrations may cause interference phenomena such as shaking and smearing to the collected images, which have a great impact on image quality. If the industrial camera is operated under the condition of long-term vibration, the fixed position of the camera will shift, which will affect the final collection effect and cause certain damage to the camera. Therefore, a stable mechanical structure is required to relatively fix the position of the camera.

(2) Device position adjustment: In order to collect images more accurately and adjust the position of the industrial camera in a timely and rapid manner, the detection device also needs to have adjustment functions such as up and down, left and right, and rotation.

(3) Camera calibration: By changing the pose of the camera, the internal parameters of the camera can be easily obtained from multiple image data obtained by Zhang Zhengyou’s calibration method, so as to realize the online calibration of the product. Finally, when the positions of the camera and the product to be tested are not changing, the external parameters of the camera can be obtained according to the position of the product.

(4) On-line detection and measurement: collect images to the computer through cameras, data interfaces, etc., and realize product surface quality detection (such as defects, characters, etc.) and shape measurement in the image processing module of the computer, and can feedback the detection and measurement results in real time To the control system, the control system performs adaptive calibration and adjustment to the vision system according to the results to meet the best measurement effect.

According to above-mentioned inventive concept, the present invention adopts following technical scheme:

A self-adaptive calibration visual online detection device, including a base, an industrial camera, an industrial computer, a conveyor belt and an object to be measured, is characterized in that:

1) The base is fixed on the ground, next to the conveyor belt, on which objects to be detected are placed, and an electric rotary table A is installed on the base;

2) Install a motor-screw device A vertically on the electric turntable A, wherein the lead screw A is driven by a stepping motor A, and the lead screw A is screw-fitted with an electric turntable B;

3) Install a motor-screw device B horizontally on the electric rotary table B, wherein the screw B is driven by a stepping motor B, and the screw B is screw-fitted with an electric rotary table C;

4) The industrial camera is installed on the electric rotary table C through a camera mounting plate;

5) The electric turntable A, electric turntable B, stepper motor A, electric turntable C and stepper motor B are all connected to the computer via a controller.

The visual on-line detection method of self-adaptive calibration uses the above-mentioned device to operate, and the specific steps are: before calibration, put a calibration plate on the conveyor belt; then, the set program controls the industrial camera to return to the initial position, that is, the camera lens Right below; then, start the automatic calibration of the camera; the set calibration program controls the industrial camera to move along the Y-axis and Z-axis, and to adjust the camera's rotation around the X-axis, Y-axis, and Z-axis respectively. Attitude; adjust the camera attitude in a small range through the set program to obtain multiple different pictures of the calibration board; then, process the photos by the set calibration program to obtain the internal reference of the camera; after the calibration is completed, the computer will automatically After entering the detection program interface, the integration of calibration and detection is realized; after image acquisition, image preprocessing, image segmentation, and morphological processing operations are completed, detection and measurement are performed automatically. If the detection results are inconsistent with the standard sample parameters or the detection effect is not good, Then it will return to the calibration interface to calibrate, thus realizing the intelligent adjustment function between calibration and detection; after the calibration is completed, it will return to the detection interface to work again.

Compared with the prior art, the present invention has the following prominent advantages:

The device has simple and practical structure, convenient operation, high calibration precision, and is stable and reliable. The calibration process is fully automatic without manual intervention, eliminating the interference and errors caused by manual factors, and effectively improving the calibration accuracy. At the same time, even non-professionals can operate it.

Description of drawings

Figure 1 is a diagram of a visual online inspection device for adaptive calibration.

Fig. 2 is a structural block diagram of an adaptively calibrated visual online inspection device.

Fig. 3 is a flowchart of a visual online detection method for adaptive calibration.

detailed description

Preferred embodiments of the present invention are described in detail as follows in conjunction with accompanying drawings:

Embodiment one:

Referring to Fig. 1, this self-adaptive calibration visual online detection device includes a base (1), an industrial camera (10), an industrial computer (15), a conveyor belt (13) and an object to be measured (12), and is characterized in that:

1) The base (1) is fixed on the ground, next to the conveyor belt (13), on which the object (12) to be detected is placed, and an electric rotary table A (2) is installed on the base (1);

2) Install a motor-screw device A vertically on the electric rotary table A (2), wherein the screw A is driven by a stepping motor A (5), and the screw A is connected with an electric rotary table B (6 ) screw fit;

3) Install a motor-screw device B horizontally on the electric rotary table B (6), wherein the screw B is driven by a stepping motor B (11), and the screw B is connected with an electric rotary table C (8) screw fit;

4) The industrial camera (10) is installed on the electric rotary table C (8) through the camera mounting plate (9);

5) The electric rotary table A (2), electric rotary table B (6), stepper motor A (5), electric rotary table C (8) and stepper motor B (11) are all controlled by a controller (14 ) connected to said computer (15).

Embodiment 2: This embodiment is basically the same as Embodiment 1, and the special features are as follows:

The base (1) and the electric rotary table A (2), between the electric rotary table A (2) and a support plate (3) are connected by screws; the support plate (3) and a screw table A (4) are connected by bolts. Between the screw table A (4) and the electric rotary table B (6), between the electric rotary table B (6) and a screw table B (7), between the screw table B (7) and the electric Between the rotating table C (8), between the electric rotating table C (8) and a camera mounting plate (9), and between the camera mounting plate (9) and the industrial camera (10) are connected by screws; the steps Both the stepper motor A (5) and the stepper motor B (11) are screwed to the end of the lead screw table (4, 7) through a mounting plate; the stepper motor A (5), stepper motor B (11), The electric turntable A (2), electric turntable B (6) and electric turntable C (8) are all connected to the controller (14) through the RS232 serial port, and the controller (14) is also connected to the computer (15) through the serial port. Through 3 electric rotating tables (2, 6, 8) and 2 stepping motor-screw devices, the free movement of the industrial camera (10) along the Y-axis and Z-axis directions, and around the X-axis and Y-axis respectively Axis and Z axis rotate, so that the industrial camera (10) has 5 degrees of freedom.

Embodiment three:

Referring to Fig. 1 and Fig. 2, this self-adaptive calibration visual on-line detection system consists of three parts: transportation device, detection device and control device. The transportation device is composed of a conveyor belt (13) and an object to be measured (12), and the control device is composed of a controller (14) and a computer (15). The detection device consists of a base (1), an electric rotary table A (2), a support plate (3), a screw table A (4), a stepping motor A (5), an electric rotary table B (6), and a screw table B (7), electric rotary table C (8), camera mounting plate (9), industrial camera (10), stepping motor B (11), object to be measured (12), conveyor belt (13), controller (14 ), a computer (15) constitutes. The base (1) and the electric turntable A (2), and between the electric turntable A (2) and the support plate (3) are connected by screws. The support plate (3) is connected with the lead screw table A (4) by bolts. Between the screw table A (4) and the electric rotary table B (6), between the electric rotary table B (6) and the screw table B (7), between the screw table B (7) and the electric rotary table C (8 ), between the electric rotary table C (8) and the camera mounting plate (9), and between the camera mounting plate (9) and the industrial camera (10) are connected by screws. Both the motor A and the motor B are screwed to the end of the lead screw table through the mounting plate. Stepping motor A (5), stepping motor B (11), electric rotary table A (2), electric rotary table B (6) and electric rotary table C (8) are all connected to the controller through the RS232 serial port, and the controller It is also connected with the computer through the serial port.

Embodiment four:

Referring to Figure 3, the visual online detection method of adaptive calibration in this book uses the above-mentioned device for operation, and the operation steps are as follows:

Before calibration, put a calibration plate on the conveyor belt; then, the set program controls the industrial camera to return to the initial position, that is, the camera lens is facing directly downward; then, the automatic calibration of the camera starts; controlled by the set calibration program The industrial camera moves along the Y-axis and Z-axis, and rotates around the X-axis, Y-axis, and Z-axis respectively to adjust the attitude of the camera; adjust the attitude of the camera in a small range through the set program, you can get Multiple different pictures of the calibration board; then, the internal reference of the camera is obtained by processing the photos with the set calibration program; after the calibration is completed, the computer will automatically enter the detection program interface, realizing the integration of calibration and detection; complete image acquisition, image After preprocessing, image segmentation, and morphological processing, the detection and measurement will be performed automatically. If the detection result is inconsistent with the standard sample parameters or the detection effect is not good, then it will return to the calibration interface for calibration, thus realizing the balance between calibration and detection. Intelligent adjustment function; after the calibration is completed, it will return to the detection interface to work again.

Working principle of the present invention is specifically as follows:

First, the camera is fixed by the camera mounting plate (9) through threaded connections. Secondly, the stepper motor A (5) and stepper motor B (11) move the camera along the Y and Z axes (see Figure 1 for coordinate system details), and the electric rotary table A (2) and electric rotary table B (6) The electric rotating table C (8) realizes the rotation of the camera on the XOZ, YOZ, and XOY planes. Through the adjustment functions such as up and down, left and right, and rotation, the camera position can be adjusted very conveniently. Since the camera has 5 degrees of freedom (along the Y-axis, Z-axis direction, rotation around the X-axis, Y-axis, Z-axis), pose transformation is easy and fast. By controlling the stepping motors A, B and the electric rotating tables A, B, C by the controller 14, the position adjustment of the camera can be realized automatically, so as to obtain the pictures of the calibration board in different poses. The collected images are sent to the computer (15), and the calibration program of the computer's internal software (according to Zhang Zhengyou's calibration method) can realize the calibration of the camera. After the calibration is completed, the computer pose will not change, and it will automatically enter the object detection stage. Detection mainly includes operations such as image acquisition, image preprocessing, image segmentation, and morphological processing. According to the data after image detection, the measurement work can be completed. At the same time, the measurement results are fed back to the computer for adjustment.

Claims (4)

1. the vision on-line measuring device of a kind of self-adapting calibration, including base（1）, industrial camera（10）, industrial computer （15）, conveyer belt（13）And object under test（12）, it is characterised in that：

1) base（1）Ground is fixed on, positioned at conveyer belt（13）Side, conveyer belt（13）On shelve examined object（12）, Base（1）One electric rotary table A of upper installation（2）；

2) in the electric rotary table A（2）A upper vertically-mounted motor-feed screw apparatus A, wherein leading screw A is by motor A （5）Drive, and a leading screw A and electric rotary table B（6）Threaded engagement；

3) in the electric rotary table B（6）On be horizontally mounted a motor-feed screw apparatus B, wherein leading screw B is by motor B （11）Drive, and a leading screw B and electric rotary table C（8）Threaded engagement；

4) industrial camera（10）By camera installing plate（9）Installed in the electric rotary table C（8）On；

5) the electric rotary table A（2）, electric rotary table B（6）, motor A（5）, electric rotary table C（8）With stepping electricity Machine B（11）All through a controller（14）Connect the computer（15）.

2. the vision on-line measuring device of self-adapting calibration according to claim 1, it is characterised in that：The base（1） With electric rotary table A（2）Between, electric rotary table A（2）With a gripper shoe（3）Between be connected by screw；The gripper shoe （3）With a leading screw platform A（4）Between be bolted,

The leading screw platform A（4）With electric rotary table B（6）Between, electric rotary table B（6）With a leading screw platform B（7）Between, institute State leading screw platform B（7）With electric rotary table C（8）Between, electric rotary table C（8）With a camera installing plate（9）Between, the phase Machine installing plate（9）With industrial camera（10）Between be connected by screw；The motor A（5）With motor B（11）All lead to Cross an installing plate and leading screw platform（4、7）End carries out mode connects for screw；Motor A（5）, motor B（11）, electronic rotation Platform A（2）, electric rotary table B（6）And electric rotary table C（8）All by RS232 serial ports and controller（14）Connection, controller （14）With computer（15）Connect also by serial ports, by 3 electric rotary tables（2、6、8）And 2 motor-leading screw dresses Put, it is achieved that industrial camera（10）Moving freely in Y-axis, Z-direction, and rotate rotating around X-axis, Y-axis, Z axis, so as to Make industrial camera（10）Have 5 degree of freedom.

3. a kind of vision online test method of self-adapting calibration, adopts the vision of self-adapting calibration according to claim 1 On-line measuring device is operated, it is characterised in that concretely comprised the following steps：Before demarcation, in conveyer belt（13）On put one piece of scaling board （12）；Then, by the process control industry camera for having set（10）Initial position is returned to, i.e., camera lens are towards underface；Tightly Then, start the automatic Calibration of camera；By the calibrating procedure control industrial camera for setting（10）Move along Y-axis, Z-direction, with And rotating around X-axis, Y-axis, the rotary motion in Z-direction is adjusting the attitude of camera；By the program that set to camera appearance The a small range adjustment of state, is obtained multiple different pictures of scaling board；Then, process photo by the calibrating procedure for having set to obtain Internal reference to camera；Then, camera attitude keeps constant, automatically into the detection-phase of object, carries out image acquisition, image pre- Process, image segmentation, Morphological scale-space operation；After image detection measurement is finished, according to detected value, calibration result is tested, If testing result deviation is larger, computer will re-start demarcation；After the completion of re-scaling, then detected.

4. the vision online test method of self-adapting calibration according to claim 3, it is characterised in that：After the completion of demarcation, Computer（15）Will be automatically into detection program interface, it is achieved that demarcate integrated with detection；Complete image acquisition, image pre- After process, image segmentation, Morphological scale-space are operated, detection measurement is carried out automatically, if testing result is differed with master sample parameter Cause or Detection results are bad, then calibration interface will be come back to and demarcated, it is achieved thereby that demarcating and detection between Adjustment function；After the completion of demarcation, detection interface will be come back to again and be operated.