A calibration method for the parallelism difference of the optical axes of a multi-sensor photoelectric theodolite

Abstract

The present invention provides a calibration method for the optical axis parallelism difference of a multi-sensor photoelectric theodolite, which can accurately measure the optical axis parallelism difference value of the heterologous sensors of the photoelectric theodolite. By designing a special field detection device for infrared and visible light sensors and proposing two marker point positioning methods for the two types of images, the present invention realizes the calibration of the optical axis parallelism difference of the heterologous sensors of the photoelectric theodolite, which has great significance for ensuring the measurement accuracy of the photoelectric theodolite.

Description

Technical Field

[0001] The present invention belongs to the field of optical measurement and image processing, and particularly relates to a method for calibrating the optical axis parallelism difference of a multi-sensor photoelectric theodolite. Background Art

[0002] A photoelectric theodolite is an optical measurement device. When measuring small targets flying at high speeds in the air, due to the relatively long flight distance of the target, a photoelectric theodolite with both infrared and visible light sensors is generally used. When the measured target is relatively close, a visible light sensor is generally used. When the distance is relatively far, an infrared sensor is generally used. In order to ensure the light input of the lenses of each sensor, independent optical paths are generally adopted, that is, the infrared and visible light respectively have their own optical paths. In this way, during the actual installation of optical devices, the two optical paths cannot be completely parallel. And this parallelism difference is calibrated in the assembly workshop using special calibration equipment. Currently, all standard calibration methods require a professional optical frame and are completed in the assembly laboratory. However, during the road transportation and field installation of the theodolite on the shooting range, vibrations will inevitably occur, and the optical axis parallelism difference of the heterologous sensors will inevitably change. And the photoelectric theodolite is a precision angle measuring device that does not allow any errors. Therefore, the existing technology has the disadvantages of extremely demanding requirements for the external working environment and being unable to completely eliminate the parallelism error. Due to the external transportation, installation, and climate change in the test environment, the rigid structure of the photoelectric theodolite cannot be guaranteed to be completely unchanged. And due to the limitations of the external field conditions, it is impossible to install a professional optical inspection frame. Therefore, this error is difficult to avoid. Summary of the Invention

[0003] In view of this, the present invention provides a method for calibrating the optical axis parallelism difference of a multi-sensor photoelectric theodolite, which can accurately measure the optical axis parallelism difference value of the heterologous sensors of the photoelectric theodolite.

[0004] To achieve the above object, the technical solution of the present invention is as follows:

[0005] A method for calibrating the non-parallelism of the optical axes of a multi-sensor photoelectric theodolite includes the following steps: Visible light detection marks and infrared detection marks are arranged at intervals in front of the photoelectric theodolite. The marks face the photoelectric theodolite directly, and the geodetic height of the center of the mark pattern is the same as the height of the three-axis center of the photoelectric theodolite. The two marks form a certain angle with the three-axis center of the photoelectric theodolite;

[0006] The visible light detection mark image collected by the photoelectric theodolite is subjected to binarization processing. Then, the binarized image is subjected to Hough transform processing to obtain the mark center position. Two methods of obtaining the intersection of two straight lines in the mark and the center of the mark circle are used to obtain the mark center. Then, the average value of the results of the two methods is taken as the final value of the mark center position;

[0007] For the infrared detection mark, the infrared detection mark image collected by the photoelectric theodolite is binarized, and the gray center of gravity of the target area of the binarized image is taken as the infrared source center;

[0008] The infrared and visible light centers are interpreted to obtain the comprehensive angle values of these two sensors respectively.

[0009] During shooting, the center of the visible light detection mark is aligned with the center of the visible light sensor image. The infrared and visible light detection mark centers are interpreted and processed, and the interpreted value minus the actual included angle between the two optical axes is the optical axis parallelism difference.

[0010] Among them, the visible light detection mark has a black circumference on the outer edge, and the circumference is divided into four equal parts by angles, with two black and two white parts in the four equal parts, alternating black and white; the infrared detection mark is a standard halogen lamp, and the lamp surface faces the photoelectric theodolite.

[0011] Among them, the specific implementation method for obtaining the center of the mark circle is as follows:

[0012] Establish the circle equation in the image space: (x - a) 2 +(y - b) 2 = r 2 , where (a, b) is the center of the circle, r is the radius, and (x, y) is the point on the circumference; any edge point (x i , y i ) in the image space corresponds to a conical surface on the parameter space (a, b, r) after Hough transformation; the conical surfaces corresponding to all the points on the same circle in the image space intersect at a point (a0, b0, c0), and this point exactly corresponds to the center of the circle (a0, b0) and the radius r0. The above formula is transformed into the parameter equation as follows:

[0013]

[0014] where θ ∈ [0, 2π), r ∈ [R min , R max ; the parameter space is quantized to obtain a 3D accumulator array A(a, b, r), and each cube cell in the array corresponds to the discrete parameter values of (a, b, r).

[0015] When detecting a circle in an image, first calculate the gradient information of each point in the image, then find the edge points according to an appropriate threshold, and then traverse the parameters θ and r within their value ranges with their respective quantization intervals as step sizes, calculate all points (a, b) whose distance from each pixel point on the edge is r, and at the same time increment the accumulator A(a, b, r) of the corresponding (a, b, r) cubic cell by 1; after the transformation of all edge points is completed, the local peak cell of the accumulator A(a, b, r) corresponds to the circle parameters in the image space, and the center of the circle corresponding to the maximum value of the accumulator is the position of the marker center.

[0016] Among them, the specific implementation method for finding the intersection point of two straight lines in the marker is as follows:

[0017] Establish the polar coordinate formula ρ = x cosα + y sinα in the Cartesian coordinate system. Each pixel coordinate p(x, y) in the image space is known, and ρ and α are variables to be found. For each pixel coordinate, a series of (ρ, α) values are obtained by looping through the angle α. If the pixel points are on the same straight line, the corresponding (ρ, α) values in the parameter space are concentrated at one point. Finally, the two largest sets of values in the accumulative array are found by searching for the peak, which are the intersection points of the two lines of the visible light detection marker, that is, the position of the marker center.

[0018] Among them, the method of taking the gray center of the binarized target area as the infrared source center is as follows: Statistically calculate the gray center of gravity of the pixels in a square area with a side length of n pixels. The formula is Among them, x i or y i is the center of gravity coordinate of the horizontal or vertical axis, i or j is the center of gravity coordinate of the horizontal or vertical axis, and the obtained (x, y) coordinates are the position of the marker center.

[0019] Among them, the angle measurement of the photoelectric theodolite is calculated according to the following formula: A = a + x × d x , E = e + y × d y ; where, A is the combined angle value of the azimuth angle, E is the combined angle value of the elevation angle, a is the azimuth angle of the photoelectric theodolite encoder, e is the elevation angle of the photoelectric theodolite encoder, x and y are the number of pixels by which the target in the image deviates from the image center, d x , d y is the dimension value of the sensor.

[0020] Among them, the optical axis parallelism error is as follows:

[0021]

[0022] Among them, θ is the included angle between the infrared and visible light sensors, is the required parallelism error.

[0023] Among them, each visible light detection mark and infrared detection mark capture more than 30 images, and the acquisition time interval of each image is greater than 0.5 seconds. Each image is interpreted, and the average value of the interpretation results is taken as the final result.

[0024] Beneficial effects

[0025] 1. By designing a special field detection device for infrared and visible light sensors and proposing two methods for locating fiducial points in images, the present invention realizes the calibration of the optical axis parallelism difference of heterologous sensors of an optoelectronic theodolite, which is of great significance for ensuring the measurement accuracy of the optoelectronic theodolite.

[0026] 2. The present invention can calibrate the parallelism difference of heterologous sensors at any time in the field environment before the test. The scheme is simple, easy to implement, and has high calibration accuracy, which is of great significance for ensuring the measurement accuracy of the optoelectronic theodolite.

[0027] 3. The outer edge of the visible light detection mark of the present invention is a black circumference, which is divided into four equal parts by angles, two black and two white in the four equal parts, black and white alternating. The center point of the mark is at the same geodetic height as the center of the three axes of the optoelectronic theodolite, and the mark center is positioned relatively accurately.

[0028] 4. To eliminate the influence of atmospheric disturbance, each visible light detection mark and infrared detection mark in the present invention capture more than 30 images, and the acquisition time interval of each image is greater than 0.5 seconds. Each image is interpreted, and the average value of the interpretation results is taken as the final result. Description of the drawings

[0029] Figure 1 It is a schematic diagram of the principle of the measurement method of the present invention.

[0030] Figure 2 It is a specific measurement schematic diagram of an embodiment of the present invention Detailed implementation manners

[0031] The following are specific embodiments of the present invention with reference to the drawings and detailed descriptions.

[0032] The present invention provides a method for calibrating the optical axis parallelism difference of a multi-sensor optoelectronic theodolite. By using a special field detection device for infrared and visible light sensors, two methods for locating fiducial points in images are respectively proposed, which are as follows:

[0033] Visible light detection marks and infrared detection marks are respectively set at different distances in front of the optoelectronic theodolite. Since the visible light is the imaging of the sun's reflection and the mark recognition is the recognition of corner points, it is greatly restricted by the imaging conditions. Therefore, the distance of the visible light detection mark from the optoelectronic theodolite cannot be too far, and generally 1000 meters is more appropriate. The infrared detection mark is generally an infrared source and uses centroid recognition. When the distance from the optoelectronic theodolite is relatively close, the image noise has a greater impact. Therefore, the general mark distance is about 3000 meters. The visible light detection mark and the infrared detection mark form a certain angle with the three-axis center of the optoelectronic theodolite, and the angle ensures that the two marks are imaged simultaneously within the frame of the optoelectronic theodolite.

[0034] In this embodiment, the outer edge of the visible light detection mark is a black circle with a radius of one meter, the width of the circle is 0.2 meters, the circle in the circle is divided into four equal parts by angles, two black and two white in the four equal parts, black and white alternating. The center point of the mark is at the same geodetic height as the three-axis center of the optoelectronic theodolite, and the mark center is positioned relatively accurately. The positioning accuracy requirement is relatively high. The visible light detection mark center in this embodiment is positioned relatively accurately. Therefore, generally, this type of visible light detection mark is used. Extracting the center of the circle of the visible light detection mark is the mark center. In order to improve the extraction accuracy, the image collected by the optoelectronic theodolite is binarized, and then, the binarized image is subjected to Hough transform processing to extract two vertical lines of the mark, and the intersection point of the lines is the mark center. The average value of the mark center values extracted by the two extraction methods is the final value of the mark center.

[0035] Among them, the specific implementation method for obtaining the center of the mark circle is as follows:

[0036] Establish the circle equation in the image space: (x - a) 2 +(y - b) 2 =r 2 , where (a, b) is the center of the circle, r is the radius, and (x, y) are the points on the circumference. Any edge point (x i , y i ) in the image space corresponds to a conical surface in the parameter space (a, b, r) after Hough transform. All the points on the same circle in the image space correspond to the conical surfaces in the parameter space that intersect at a point (a0, b0, c0), and this point exactly corresponds to the center of the circle (a0, b0) and the radius r0.

[0037] Convert the above formula into a parametric equation as follows:

[0038]

[0039] where θ ∈ [0, 2π), r ∈ [R min , R max. Appropriately quantize the parameter space to obtain a 3D accumulator array A(a, b, r). Each small cube cell in the array corresponds to the discrete parameter values of (a, b, r). When detecting circles in the image, first calculate the gradient information of each point in the image, then find the edge points according to an appropriate threshold, and then traverse the values of the parameters θ and r within their respective value ranges with their respective quantization intervals as steps, calculate all points (a, b) at a distance of r from each pixel point on the edge, and at the same time increment the accumulator A(a, b, r) corresponding to the (a, b, r) cube cell by 1. After completing the transformation for all edge points, the local peak cell of the accumulator A(a, b, r) corresponds to the circle parameters in the image space. The center of the circle corresponding to the maximum value of the accumulator is the position of the marker center.

[0040] The specific implementation method for obtaining the intersection point of two straight lines in the marker is as follows:

[0041] Establish the polar coordinate formula ρ = x cosα + y sinα in the Cartesian coordinate system. The pixel coordinates p(x, y) of each point in the image space are known, and ρ and α are the variables to be searched. For each pixel coordinate, a series of (ρ, α) values are obtained by looping through the angle α. If the pixel points are on the same straight line, the corresponding (ρ, α) values in the parameter space are concentrated at one point. Finally, the two largest sets of values in the accumulative array are found by searching for the peak, which are the intersection points of the two lines of the visible light detection marker, that is, the position of the marker center.

[0042] The infrared detection marker uses an infrared light source. In this embodiment, a standard halogen lamp is used, and the lamp surface faces the photoelectric theodolite. When extracting the image collected by the photoelectric theodolite, first perform binarization processing, and secondly, take the gray center of the binarized target area as the infrared source center. The processing method is as follows: Statistically calculate the gray center of gravity of the pixels within a square area with a side length of n pixels. The formula is as follows:

[0043]

[0044]

[0045] Among them, x i or y i is the center of gravity coordinate of the horizontal or vertical axis, i or j is the center of gravity coordinate of the horizontal or vertical axis, and the obtained (x, y) coordinates are the position of the marker center.

[0046] To eliminate the influence of atmospheric disturbance, more than 30 images are taken for each visible light detection marker and infrared detection marker, and the acquisition time interval of each image is greater than 0.5 seconds. Each image is interpreted, and the average value of the interpretation results is taken as the final result. Among them, the angle measurement of the photoelectric theodolite can be calculated according to the following formula:

[0047] A = a + x × d x

[0048] E = e + y × d y

[0049] Wherein, A is the azimuth comprehensive angle value, E is the elevation comprehensive angle value, a is the azimuth angle of the photoelectric theodolite encoder, e is the elevation angle of the photoelectric theodolite encoder, x and y are the number of pixels by which the target in the image deviates from the image center, and d x , d y is the dimension value of the sensor. The infrared and visible light centers are interpreted to obtain the comprehensive angle values of these two sensors respectively.

[0050] Currently, the standard method for detecting the flat phase difference is the method of using a collimator in the laboratory. The present invention is a field detection method for the photoelectric theodolite during actual operation. During shooting, the center of the visible light detection mark is aligned with the image center of the visible light sensor as much as possible. The infrared and visible light detection mark centers are interpreted and processed, and the interpreted value minus the actual included angle between the two optical axes is the optical axis parallel difference, as shown in the following formula:

[0051]

[0052] Wherein, θ is the included angle between the infrared and visible light sensors, is the required parallel difference.

[0053] In summary, the above are only the preferred embodiments of the present invention and are not used to limit the protection scope of the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A calibration method for the non - parallelism of the optical axes of a multi - sensor photoelectric theodolite, characterized in that, Including the following steps: Set a visible light detection mark and an infrared detection mark at intervals in front of the photoelectric theodolite. The marks face the photoelectric theodolite directly. The geodetic height of the center of the mark pattern is the same as the height of the three-axis center of the photoelectric theodolite. The two marks form a certain angle with the three-axis center of the photoelectric theodolite; Perform binarization processing on the visible light detection mark image collected by the photoelectric theodolite. Then, perform Hought transform processing on the binarized image to obtain the center position of the mark. Two methods, namely, finding the intersection point of two straight lines in the mark and the center of the mark circle, are used to find the center of the mark. Then, take the average value of the results of the two methods as the final value of the center position of the mark; For the infrared detection mark, perform binarization processing on the infrared detection mark image collected by the photoelectric theodolite, and take the gray center of gravity of the target area of the binarized image as the infrared source center; The specific implementation method for finding the intersection point of two straight lines in the mark is as follows: Establish a polar coordinate formula ρ = x cosα + y sinα in the Cartesian coordinate system. The pixel coordinates p(x, y) in the image space are known, and ρ and α are variables to be found. For each pixel coordinate, a series of (ρ, α) values are obtained by looping through the angle α. If the pixel points are on the same straight line, the corresponding (ρ, α) values in the parameter space are concentrated at one point. Finally, by searching for the peak, the two largest values in the cumulative array are found, which are the intersection points of the two lines of the visible light detection mark, that is, the center position of the mark; Among them, the visible light detection mark and the infrared detection mark are respectively set at different distances in front of the photoelectric theodolite. The visible light detection mark is 1000 meters away from the photoelectric theodolite, and the infrared detection mark is an infrared source. Centroid recognition is used, and it is 3000 meters away from the photoelectric theodolite. The outer edge of the visible light detection mark is a black circle with a radius of one meter, and the width of the circle is 0.2 meters. The circle in the circle is divided into four equal parts by angles, with two black and two white, alternating black and white. The center point of the mark is at the same geodetic height as the three-axis center of the photoelectric theodolite. The infrared detection mark uses an infrared light source, and the lamp surface faces the photoelectric theodolite directly; When shooting, align the center of the visible light detection mark with the center of the visible light sensor image, and perform interpretation processing on the centers of the infrared and visible light detection marks. The interpretation value minus the actual included angle between the two optical axes is the parallel difference of the optical axes, as shown in the following formula: Among them, θ is the included angle between the infrared and visible light sensors, A is the azimuth angle spatial pointing angle value, and a is the azimuth angle of the photoelectric theodolite encoder. is the required parallel error.

2. The calibration method according to claim 1, wherein The angle measurement of the optoelectronic theodolite is calculated according to the following formula: A = a + x × d x , E = e + y × d y ; where E is the spatial pointing angle value of the pitch angle, a is the azimuth angle of the optoelectronic theodolite encoder, e is the pitch angle of the optoelectronic theodolite encoder, x and y are the number of pixels by which the target in the image deviates from the image center, d x , d y is the dimension value of the sensor.

3. The calibration method according to claim 1, characterized in that Each visible light detection mark and infrared detection mark are photographed with more than 30 frames of images, and the time interval between each frame of image acquisition is greater than 0.5 seconds. Perform interpretation on each frame of image, and take the average value of the interpretation results as the final result.

Images