Method, system and device for compensating image motion of two-dimensional optics based on PSD

Abstract

The method, system and device for compensating image motion in two-dimensional optics based on PSD for rapid translation belong to the field of optical image motion compensation technology, solve the problem of large deviation in the position measurement results of the existing optical image motion compensation system based on PSD, and can replace the rapid reflection image motion compensation method. The method of the present invention includes: a semiconductor laser module emits a light beam, which is vertically irradiated onto the photosensitive surface of the PSD device of the corresponding axis, and the photoelectric analog signal generated by the light spot on the PSD device of the axis; the photoelectric analog signal is amplified and converted into a digital signal; according to the converted digital signal, position resolution, filtering and fitting correction are performed to obtain the final light spot position information; an integral operation is performed on the residual value of the external frame gyro, the difference in angular velocity is converted into an angle difference, the angle difference is converted into a radian difference, and then multiplied by the focal length value of the imaging system to obtain the linear displacement to be compensated; the lens is controlled to perform two-dimensional high-speed linear translation along the X and Y axes to compensate for the image motion.

Description

Technical Field

[0001] The present application relates to the technical field of optical image motion compensation, and in particular to PSD-based two-dimensional optical rapid translational image motion compensation. Background Art

[0002] When a telephoto camera is used in an airborne or shipborne environment, the image at the time of exposure will move on the focal plane due to the change in the carrier's attitude, atmospheric airflow disturbances and vibrations, resulting in blurred images and affecting resolution and target recognition. This is especially obvious in cameras with long focal lengths, small CCD pixel sizes, and long exposure times. The aerial photography industry standard usually requires that the image shift at the time of exposure should not exceed 1 pixel. To solve this problem, telephoto cameras usually use an external frame gyro mechanism to perform rough motion compensation for disturbances, and then use the high bandwidth and fast response characteristics of the fast reflector (FSM) to perform fine compensation for the rough compensated gyro residuals.

[0003] The premise of using a fast reflector (FSM) is that a long focal length optical lens artificially introduces a parallel light path during optical design, which increases the difficulty of optical system design and occupies the structural space size. At the same time, the probe and signal processing board of the contactless eddy current position sensor of the fast reflector are large in size, the coaxial cable used for probe signal transmission also occupies space size, and the sensor is expensive.

[0004] PSD is a photosensitive electronic component with a response time of microseconds, a position resolution of less than 1μm, and a position accuracy of μm. The output analog signal can achieve a sampling frequency of hundreds of KHz, and the position measurement result is independent of the spot shape, only related to the center of gravity of the spot, and the cost is low. It is widely used in precision instrument measurement. However, PSD signal extraction belongs to weak signal extraction, with much signal noise interference, and high requirements for signal extraction, amplification, solution, filtering and calibration, otherwise it will lead to deviations in position measurement results. Summary of the invention

[0005] The purpose of the present invention is to solve the problem of large deviation in position measurement results of existing PSD-based optical image motion compensation systems, and to replace the fast reflector mirror (FSM) image motion compensation method, and to provide a PSD-based two-dimensional optical fast translational image motion compensation method, system and device.

[0006] The present invention is implemented by the following technical solutions. On the one hand, the present invention provides a method for compensating for two-dimensional optical fast translational image motion based on PSD. The method is based on a non-contact photoelectric displacement measurement device for translational image motion compensation. The non-contact photoelectric displacement measurement device includes a semiconductor laser module and a one-dimensional PSD device.

[0007] There are two one-dimensional PSD devices, one for the x-axis and one for the y-axis, which are arranged parallel to the x-axis and the y-axis in the direction of lens movement and do not move with the lens;

[0008] There are two semiconductor laser modules, which are used in conjunction with the one-dimensional PSD devices of the x-axis and the y-axis respectively, and move with the lens. When the device is stationary, the light beams emitted by the two semiconductor laser modules are vertically irradiated to the center of the photosensitive surface of the corresponding one-dimensional PSD device;

[0009] The method specifically comprises:

[0010] Step 1: The semiconductor laser module emits a light beam, which is vertically irradiated onto the photosensitive surface of the one-dimensional PSD device of the corresponding axis, and the light spot generates a photoelectric analog signal on the one-dimensional PSD device of the axis;

[0011] Step 2, amplify the photoelectric analog signal and convert it into a digital signal;

[0012] Step 3: According to the converted digital signal, position calculation, filtering and fitting correction are performed to obtain the final spot position information;

[0013] Step 4: Integrate the residual value of the external frame gyro to convert the difference in angular velocity into an angle difference, convert the angle difference into a radian difference, and then multiply it by the focal length of the imaging system to obtain the linear displacement to be compensated. ;

[0014] Step 5: Change the linear displacement of a certain axis As a given value, the spot position information after the axis is fitted As the feedback value, the difference between the feedback value and the given value is compared in real time. According to the positive and negative and size of the difference, the lens is controlled to make two-dimensional high-speed linear translation motion along the X and Y axes to compensate for the image shift. When the difference is zero, the image shift compensation is completed and the lens stops moving.

[0015] Furthermore, in step 3, the position solution specifically includes:

[0016] The spot position information of a one-dimensional PSD device on a certain axis is obtained by the spot offset position relative to the center position of the PSD on that axis. The calculation formula is:

[0017]

[0018] in, and It is the AD code value of the photocurrent signal converted from the electrodes at both ends of the corresponding one-dimensional PSD. S is half of the effective length of the effective photosensitive surface of the PSD device on this axis. The calculated offset relative to the PSD center position of the axis is the spot position information.

[0019] Furthermore, in step 3, filtering is performed on the light spot position information, and the filtering specifically includes:

[0020] Use IIR low-pass filter to iteratively remove The high-frequency noise in the spot position information, the iterative process is as follows:

[0021] (1)

[0022] (2)

[0023] (3)

[0024] in, is the filter coefficient, 0< <1, is the filtered position signal, The old spot position value information used in the iterative calculation, The calculation is to iterate the new spot position value information, and the iterative process is to repeat formula (1), (2), and (3) in sequence.

[0025] Furthermore, in step 3, the fitting correction is the filtered position signal The generated value is obtained after the high-precision laser position sensor calibrates and corrects N positions, specifically:

[0026] A metal cube with a receiving surface perpendicular to the directions of the X and Y axes is fixed on the lens, and the cube moves together with the lens;

[0027] When the voice coil motor of a certain axis drives the lens to move the cube, the actual displacement of the cube is measured by the laser position sensor placed parallel to the axis to correct and fit the spot position information solved by PSD. The entire travel range of a certain axis is divided into N equal parts, and the step length of each part is the total travel divided by N distance, that is, each travel moves in sequence according to the spot position value. , , … , among which The position is the zero position of the measurement starting point, and the high-precision laser position sensor values ​​of the corresponding N positions are collected at the same time , , … ,After obtaining two sets of numbers, the spot position information finally generated after fitting is obtained.

[0028] Furthermore, in step 3, the fitting correction formula is:

[0029]

[0030] in, The final spot position information generated after fitting is used as position feedback of the control system. is the spot position information after filtering, and A, B, and C are binomial fitting parameters.

[0031] Furthermore, in step 4, the linear displacement of a certain axis The calculation formula is:

[0032]

[0033] in, Give a given value of the inertial space angular velocity for a certain axis, in degrees / s. is the actual feedback angular velocity value of the gyro of this axis, in degrees / s. To control the control cycle of the system, is the focal length of the camera, and 57.3 is the angle radian conversion factor.

[0034] Furthermore, the length of the photosensitive surface of the one-dimensional PSD device is not less than the effective stroke of the lens on the corresponding axis.

[0035] In a second aspect, the present invention provides a two-dimensional optical fast translational image motion compensation system based on PSD, the system comprising a contactless photoelectric displacement measurement device, a PSD sensor signal processing board, a drive control board and a push-pull voice coil motor;

[0036] The non-contact photoelectric displacement measuring device comprises a semiconductor laser module and a one-dimensional PSD device;

[0037] There are two one-dimensional PSD devices, which are divided into x-axis and y-axis one-dimensional PSD devices, which are arranged in parallel to the x-axis and y-axis in the direction of lens movement and do not move with the lens;

[0038] There are two semiconductor laser modules, which are used in conjunction with the one-dimensional PSD devices of the x-axis and the y-axis, respectively, and move with the lens. The setting position is that when the device is stationary, the light beams emitted by the two semiconductor laser modules are vertically irradiated to the center position of the photosensitive surface of the corresponding one-dimensional PSD device, and the light spot generates a photoelectric analog signal on the PSD, and the signal can be used to calculate the position information of the light spot relative to the center position of the PSD photosensitive surface;

[0039] There are two PSD sensor signal processing boards, one for the x-axis and the other for the y-axis, which are used to amplify the x-axis and y-axis light position analog signals, convert them into digital signals and send them to the drive control board;

[0040] The drive control board is used to perform spot position calculation, filtering and fitting correction according to the digital signals after x-axis and y-axis conversion, obtain the final x-axis and y-axis spot position signals, and drive the push-pull voice coil motor to drive the lens to perform two-dimensional linear motion;

[0041] The number of push-pull voice coil motors is 4, including 2 on the x-axis and 2 on the y-axis, which are arranged at both ends of the lens with the lens as the center and are parallel to the x-axis and y-axis moving directions of the lens respectively.

[0042] In a third aspect, the present invention provides a computer device, comprising a memory and a processor, wherein the memory stores a computer program, and when the processor runs the computer program stored in the memory, the steps of a two-dimensional optical rapid translational image motion compensation method based on PSD as described above are executed.

[0043] In a fourth aspect, the present invention provides a computer-readable storage medium, wherein a plurality of computer instructions are stored in the computer-readable storage medium, and the plurality of computer instructions are used to enable a computer to execute a two-dimensional optical rapid translational image motion compensation method based on PSD as described above.

[0044] Beneficial effects of the present invention:

[0045] PSD stands for Position Sensitive detector, which is a photoelectric device used in combination with a light source to detect the precise position of a light spot. The principle of the present invention is to use PSD as a position detection element to control the optical lens to quickly translate along the X and Y axes to compensate for image shift. When a long-focal-length optical camera is imaging in an airborne or shipborne environment, the image will move on the focal plane due to the shaking of the carrier, wind disturbance, and carrier vibration during camera exposure, resulting in blurred imaging, which is particularly obvious under long focal lengths or long exposure times. The solution is usually to use an external frame gyro mechanism to perform rough motion compensation for X and Y axis disturbances, and then use a fast reflector (FSM) to perform fine compensation for the rough compensated X and Y axis gyro residuals. The introduction of a fast reflector requires the optical lens to lead out a parallel optical path during design, which increases the difficulty of optical design and complicates the optical path; at the same time, the non-contact eddy current position sensor and signal processing board of the fast reflector are large in size and high in cost.

[0046] The present invention utilizes a PSD device and a semiconductor laser module to form a contactless position sensor. According to the residual value of the outer frame gyroscope, the push-pull voice coil motor is controlled to drive the lens to move two-dimensionally quickly linearly to compensate for image shift. Compared with the traditional fast reflector design method, it has two advantages: 1) The high-speed linear lens translation movement replaces the rotational movement of the reflector, and there is no need to specially design a parallel optical path, and the optical design is simple; 2) The PSD position sensor replaces the eddy current sensor, which is low in cost, small in size, and compact in overall structure.

[0047] The present invention is suitable for a secondary image motion compensation system for a long focal length optoelectronic imaging device with strict structural size requirements and simple optical design. Due to the influence of disturbances such as carrier vibration, atmospheric disturbances and attitude changes, the gyro compensation mechanism cannot completely compensate for the image motion caused by the disturbance, and a secondary image motion compensation system needs to be introduced to further stabilize the image. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to more clearly illustrate the technical solution of the present application, the drawings required for use in the embodiments are briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0049] Figure 1 A block diagram of a two-dimensional optical fast translational image shifting method based on PSD of the present invention;

[0050] Figure 2 It is a block diagram of the arrangement of the components of the system of the present invention;

[0051] Figure 3 It is a schematic diagram of the fitting correction of the laser displacement sensor of the present invention. DETAILED DESCRIPTION

[0052] Embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present invention, and should not be construed as limiting the present invention.

[0053] Embodiment 1: A two-dimensional optical fast translational image motion compensation method based on PSD, wherein the method is based on a non-contact photoelectric displacement measurement device for translational image motion compensation, wherein the non-contact photoelectric displacement measurement device comprises a semiconductor laser module 1 and a one-dimensional PSD device 2;

[0054] There are two one-dimensional PSD devices 2, one for the x-axis and one for the y-axis, which are arranged parallel to the x-axis and the y-axis in the direction of lens movement and do not move with the lens;

[0055] There are two semiconductor laser modules 1, which are used in conjunction with the one-dimensional PSD devices 2 of the x and y axes respectively, and move with the lens. The setting position is that when the device is stationary, the light beams emitted by the two semiconductor laser modules 1 are vertically irradiated to the center position of the photosensitive surface of the corresponding one-dimensional PSD device 2, and the light spot generates a photoelectric analog signal on the PSD, and the position information of the light spot relative to the center of the photosensitive surface of the PSD of the axis can be solved by using the signal;

[0056] The method specifically comprises:

[0057] Step 1: The semiconductor laser module 1 emits a light beam, which is vertically irradiated onto the photosensitive surface of the one-dimensional PSD device 2 of the corresponding axis, and the light spot generates a photoelectric analog signal on the PSD device of the axis;

[0058] Step 2: amplify the photoelectric analog signal and convert it into a digital signal on the PSD sensor signal processing board 3;

[0059] Step 3: In the drive control board 4, position calculation, filtering and fitting correction are performed according to the converted digital signal to obtain the final x- and y-axis spot position information;

[0060] Step 4. Use the known imaging system image motion compensation formula to calculate the image motion value that needs to be compensated, that is, the linear displacement that the lens needs to move on the X and Y axes. First, perform an integral operation on the external frame gyro residual value to convert the angular velocity difference into an angle difference. The calculation method is to subtract the gyro feedback velocity value from the external frame inertial space velocity given value, and then multiply it by the control time to integrate to get the angle difference; then use the relationship formula that the arc length of a circle is equal to the radian multiplied by the diameter to convert the calculated angle difference into a radian difference, and then multiply it by the focal length of the imaging system to get the linear displacement that needs to be compensated. The calculation method for the x-axis and y-axis is the same. The linear displacement of the image motion of a certain axis The calculation formula is:

[0061]

[0062] in Give a given value of the inertial space angular velocity for a certain axis, in degrees / s. is the actual feedback angular velocity value of the gyro of this axis, in degrees / s. To control the control cycle of the system, is the focal length of the camera, 57.3 is the angle radian conversion coefficient, and the calculation result converts the angular rotation required for image shift compensation into linear displacement .

[0063] Step 5: Image displacement linear displacement of the outer frame compensation of one axis As a given value, the axis is fitted The light spot position information is used as the feedback value, and the feedback value is compared with the given value in real time. According to the result of the comparison, the output of the driver is adjusted to control the operation of the voice coil motor 5, driving the lens to make a two-dimensional high-speed linear translation motion along the X and Y axes to compensate for the image shift. That is, the difference between the feedback value and the given value is compared in real time, and the lens is controlled to make a two-dimensional high-speed linear translation motion along the X and Y axes according to the positive and negative and size of the difference to compensate for the image shift. When the difference is zero, the image shift compensation is completed and the lens stops moving.

[0064] This embodiment provides a two-dimensional optical fast translational image motion compensation method based on PSD, which is used to reduce the image motion generated when imaging with a long focal length camera, and uses the two-dimensional linear motion of the lens to replace the two-dimensional rotation of the fast reflector FSM. Compared with the traditional design method of the fast reflector, it uses the high-speed linear translation of the lens to replace the rotational motion of the reflector, does not require the special design of the parallel optical path, and has a simple optical design; the PSD position sensor replaces the eddy current sensor, which is low-cost, small in size, and compact in overall structure.

[0065] Embodiment 2: This embodiment further limits the two-dimensional optical fast translational image motion compensation method based on PSD as described above. In this embodiment, step 3 is further limited, specifically including:

[0066] In step 3, the position solution specifically includes:

[0067] For the spot position information of the one-dimensional PSD device 2 of a certain axis, the calculation formula is obtained by the spot offset position relative to the center position of the PSD of the axis:

[0068]

[0069] in, and It is the AD code value of the photocurrent signal converted from the electrodes at both ends of the corresponding one-dimensional PSD. S is half of the effective length of the effective photosensitive surface of the PSD device on this axis. The calculated offset relative to the center position of the PSD of the axis is the spot position information. At this time, the spot position information needs to be filtered and fitted to further improve the accuracy.

[0070] In this embodiment, the light spot position information is solved by calculating the voltage code value output by the electrodes at both ends of the one-dimensional PSD and obtaining the offset of the light spot relative to the center point of the one-dimensional PSD on the axis.

[0071] Implementation method 3: This implementation method further limits the two-dimensional optical fast translational image motion compensation method based on PSD as described above. In this implementation method, step 3 is further limited, specifically including:

[0072] In step 3, the light spot position information needs to be filtered, and the filtering specifically includes:

[0073] Use IIR low-pass filter to iteratively remove The high-frequency noise in the spot position information, the iterative process is as follows:

[0074] (1)

[0075] (2)

[0076] (3)

[0077] in, is the filter coefficient, 0< <1, is the filtered position signal, The old spot position value information used in the iterative calculation, The calculation is to iterate the new spot position value information, and the iterative process is to repeat formula (1), (2), and (3) in sequence.

[0078] In this implementation, an IIR low-pass filter is used to filter out high-frequency noise in the system, the amount of calculation occupies less MCU resources, and it is easy to implement in a digital control system.

[0079] Embodiment 4: This embodiment further limits the two-dimensional optical fast translational image motion compensation method based on PSD as described above. In this embodiment, step 3 is further limited, specifically including:

[0080] In step 3, the fitting correction is the filtered position signal The generated value is obtained after the high-precision laser position sensor calibrates and corrects N positions, specifically:

[0081] A metal cube with a receiving surface perpendicular to the directions of the X and Y axes is fixed on the lens, and the cube moves together with the lens;

[0082] When the voice coil motor of a certain axis drives the lens to move the cube, the actual displacement of the cube is measured by the laser position sensor placed parallel to the axis to correct and fit the spot position information solved by PSD. The entire travel range of a certain axis is divided into N equal parts, and the step length of each part is the total travel divided by N distance, that is, each travel moves in sequence according to the spot position value. , , … , among which The position is the zero position of the measurement starting point, and the high-precision laser position sensor values ​​of the corresponding N positions are collected at the same time , , … , after obtaining the two sets of numbers, the spot position information finally generated after fitting is obtained.

[0083] The function of the fitting correction in this embodiment is to improve the accuracy of the light spot position, thereby improving the accuracy of image motion compensation.

[0084] Embodiment 5: This embodiment further limits the two-dimensional optical fast translational image motion compensation method based on PSD as described above. In this embodiment, step 3 is further limited, specifically including:

[0085] In step 3, the filtered spot position information needs to be fitted and corrected, and the fitting correction formula specifically includes:

[0086]

[0087] in, The final spot position information generated after fitting is used as position feedback of the control system. is the spot position information after filtering, and A, B, and C are binomial fitting parameters.

[0088] In this implementation manner, the fitting and correction of the spot position information serves to improve the accuracy of the spot position, thereby improving the accuracy of image motion compensation.

[0089] Embodiment 6: This embodiment further limits the two-dimensional optical fast translational image motion compensation method based on PSD as described above. In this embodiment, step 4 is further limited, specifically including:

[0090] In step 4, the linear displacement of a certain axis The calculation formula is:

[0091]

[0092] in, Give a given value of the inertial space angular velocity for a certain axis, in degrees / s. is the actual feedback angular velocity value of the gyro of this axis, in degrees / s. To control the control cycle of the system, is the focal length of the camera, and 57.3 is the angle radian conversion factor.

[0093] This embodiment provides a specific calculation formula for the linear displacement of a certain axis.

[0094] Embodiment 7: This embodiment further limits the two-dimensional optical fast translational image motion compensation method based on PSD as described above. In this embodiment, the one-dimensional PSD device is further limited, specifically including:

[0095] The length of the photosensitive surface of the one-dimensional PSD device 2 is not less than the effective stroke of the lens on the corresponding axis.

[0096] The purpose of this embodiment is to ensure that the light beam emitted by the semiconductor laser module 1 can form a light spot on the photosensitive surface of the one-dimensional PSD device 2 within the entire effective stroke of the lens.

[0097] Embodiment 8: This embodiment is embodiment 1 of the two-dimensional optical fast translational image motion compensation method based on PSD as described above, and specifically includes:

[0098] like Figure 1 As shown, the method specifically comprises the following steps:

[0099] Step 1: The semiconductor laser module 1 emits a light beam, which is vertically irradiated onto the photosensitive surface of the PSD device 2 of the corresponding axis, and the light spot generates a photoelectric analog signal on the PSD device of the axis;

[0100] Step 2: amplify the photoelectric analog signal and convert it into a digital signal on the PSD sensor signal processing board 3;

[0101] Step 3: In the driving control board 4, position calculation, filtering and fitting correction are performed according to the converted digital signal to obtain the final light spot position information;

[0102] Step 4. Use the known imaging system image motion compensation formula to calculate the image motion value that needs to be compensated, that is, the linear displacement that the lens needs to move on the X and Y axes. First, perform an integral operation on the external frame gyro residual value to convert the angular velocity difference into an angle difference. The calculation method is to subtract the gyro velocity feedback value from the given value of the external frame inertial space velocity, and then multiply it by the control time to integrate to get the angle difference; then use the relationship formula that the arc length of a circle is equal to the radian multiplied by the diameter to convert the calculated angle difference into a radian difference, and then multiply it by the focal length of the imaging system to convert it into the linear displacement that needs to be compensated. The calculation method for the x-axis and y-axis is the same. The image motion of a certain axis The calculation formula is:

[0103]

[0104] in Give a given value of the inertial space angular velocity for a certain axis, in degrees / s. is the actual feedback angular velocity value of the gyro of this axis, in degrees / s. To control the control cycle of the system, is the focal length of the camera, 57.3 is the angle radian conversion coefficient, and the calculation result converts the angular rotation required for image shift compensation into linear displacement .

[0105] Step 5: Image shift compensation of the outer frame of one axis As a given value, the axis after fitting The light spot position information is used as a feedback value, and the feedback value is compared with a given value in real time in the drive control board 4. According to the comparison result, the output of the driver is adjusted to control the operation of the push-pull voice coil motor 5, driving the lens to perform two-dimensional high-speed linear translation along the X and Y axes to compensate for image shift.

[0106] In step 3, the position solution specifically includes:

[0107] For the spot position information of the one-dimensional PSD device 2 of a certain axis, the calculation formula is obtained by the spot offset position relative to the center position of the PSD of the axis:

[0108]

[0109] in, and It is the AD code value of the photocurrent signal converted from the electrodes at both ends of the corresponding one-dimensional PSD. S is half of the effective length of the effective photosensitive surface of the PSD device. The calculated offset relative to the center position of the PSD of the axis is the spot position information. At this time, the spot position information needs to be filtered and fitted to further improve the accuracy.

[0110] In step 3, the light spot position information needs to be filtered, and the filtering specifically includes:

[0111] Use IIR low-pass filter to iteratively remove The high-frequency noise of the , iterative process is as follows:

[0112] (1)

[0113] (2)

[0114] (3)

[0115] in, is the filter coefficient, 0< <1, is the filtered position signal, The old spot position value information used in the iterative calculation, The calculation is to iterate the new spot position value information, and the iterative process is to repeat formula (1), (2), and (3) in sequence.

[0116] In step 3, the filtered spot position information needs to be corrected and fitted, and the fitting correction formula specifically includes:

[0117]

[0118] in, is the spot position information after filtering on a certain axis, A, B and C are binomial fitting parameters, The spot position information finally generated after fitting is used as the linear motion position feedback of the axis. The role of fitting correction is to improve the accuracy of the spot position, thereby improving the accuracy of image motion compensation.

[0119] Embodiment 9, a two-dimensional optical fast translational image motion compensation system based on PSD, specifically comprising:

[0120] like Figure 2 As shown, the system of this embodiment includes: a non-contact photoelectric displacement measuring device, a PSD sensor signal processing board 3, a drive control board 4 and a push-pull voice coil motor 5;

[0121] The non-contact photoelectric displacement measuring device comprises a semiconductor laser module 1 and a PSD device 2;

[0122] There are two PSD devices 2, which are divided into x-axis and y-axis one-dimensional PSD devices 2, which are arranged in parallel to the x-axis and y-axis in the direction of lens movement and do not move with the lens;

[0123] There are two semiconductor laser modules 1, which are used in conjunction with the one-dimensional PSD devices 2 of the x-axis and the y-axis, respectively, and move with the lens. The setting position is that when the device is stationary, the light beams emitted by the two semiconductor laser modules 1 are vertically irradiated to the center position of the photosensitive surface of the corresponding one-dimensional PSD device 2, and the light spot generates a photoelectric analog signal on the PSD, and the signal can be used to calculate the position information of the light spot relative to the center position of the PSD photosensitive surface;

[0124] There are two PSD sensor signal processing boards 3, one for the x-axis and the other for the y-axis, which are used to amplify the analog signals of the light positions of the x-axis and the y-axis, convert them into digital signals and send them to the drive control board 4;

[0125] The driving control board 4 is used to perform spot position calculation, filtering and fitting correction according to the digital signals after the x-axis and y-axis conversion, obtain the final x-axis and y-axis spot position signals, and drive the push-pull voice coil motor 5 to drive the lens to perform two-dimensional linear motion;

[0126] There are 4 push-pull voice coil motors 5, including 2 on the x-axis and 2 on the y-axis, which are arranged at both ends of the lens with the lens as the center, and are parallel to the x-axis and y-axis movement directions of the lens, respectively. When the lens moves along one axis, one voice coil motor is responsible for pushing, and one voice coil motor is responsible for pulling, and the two motors jointly drive the lens to complete the linear motion along the axis. The advantage of push-pull drive is that the force of the lens is balanced when it moves linearly, and it is easy to achieve high-frequency reciprocating motion.

[0127] The light beam emitted by the semiconductor laser module 1 is vertically irradiated onto the photosensitive surface of the one-dimensional PSD device 2 of the corresponding axis. The light spot formed generates a photoelectric analog signal on the PSD device. The analog signal is amplified by the PSD sensor signal processing board 3 and converted into a digital signal through AD conversion. The digital signal completes position resolution, filtering, and fitting correction in the drive control board 4 to generate light spot position information. At the same time, the drive control board 4 drives the push-pull voice coil motor 5 to drive the lens to move linearly in two dimensions.

[0128] The lens and semiconductor laser module 1 are installed together as the moving part, and the PSD device 2, PSD signal processing board 3 and drive control board 4 are the stationary parts. When the lens moves along one axis, the light spot will move synchronously along the axis on the PSD photosensitive surface.

[0129] A two-dimensional optical fast translation image shifting method based on PSD using the system of this embodiment includes the following steps:

[0130] Step 1, the semiconductor laser module 1 emits a light beam, which is vertically irradiated onto the photosensitive surface of the one-dimensional PSD device 2 of the corresponding axis. The photoelectric analog signal generated by the light spot on the one-dimensional PSD device 2 requires that the length of the photosensitive surface of the one-dimensional PSD device 2 is not less than the effective stroke of the lens of the axis;

[0131] Step 2: amplify the photoelectric analog signal and convert it into a digital signal on the PSD sensor signal processing board 3;

[0132] Step 3: In the drive control board 4, position calculation, filtering and fitting correction are performed according to the converted digital signal to obtain the final light spot position information;

[0133] Step 4. Use the known imaging system image motion compensation formula to calculate the image motion value that needs to be compensated, that is, the linear displacement that the lens needs to move on the X and Y axes. First, integrate the residual value of the external frame gyro to convert the difference in angular velocity into an angle difference. The method is to subtract the gyro feedback velocity value from the given value of the inertial space velocity of the external frame, and then multiply the obtained angular velocity difference by the control time for integration to obtain the angle difference; then use the relationship formula that the arc length of a circle is equal to the radian multiplied by the diameter to convert the calculated angle difference into a radian difference, and then multiply it by the focal length of the imaging system to convert it into the linear displacement that needs to be compensated. The calculation method of the x-axis and the y-axis is the same, and the image motion of one axis is The calculation formula is:

[0134]

[0135] in is the given value of the angular velocity in the inertial space of one axis, in degrees / s. is the actual feedback angular velocity value of the gyro of this axis, in degrees / s. To control the control cycle of the system, is the focal length of the camera, 57.3 is the angle radian conversion coefficient, and the calculation result converts the angular rotation required for image shift compensation into linear displacement .

[0136] Step 5: Image shift compensation of the outer frame of one axis As a given value, after fitting The spot position information is used as feedback value, and the feedback value is compared with the given value in real time. According to the comparison result, the output of the driver is adjusted to control the operation of the push-pull voice coil motor 5, driving the lens to make two-dimensional high-speed linear translation motion along the X and Y axes to compensate for image shift.

[0137] In step 3, the light spot position solution is completed in the drive control board 4, which specifically includes:

[0138] For the spot position information of the one-dimensional PSD device 2 of a certain axis, the calculation formula is obtained by the spot offset position relative to the center position of the PSD of the axis:

[0139]

[0140] in, and It is the AD code value of the photocurrent signal converted from the electrodes at both ends of the corresponding one-dimensional PSD. S is half of the effective length of the effective photosensitive surface of the PSD device. The calculated offset relative to the PSD center position is the spot position information. At this time, the spot position information needs to be filtered and fitted to further improve the accuracy.

[0141] In step 3, the light spot position information needs to be filtered, and the filtering specifically includes:

[0142] Use IIR low-pass filter to iteratively remove The high-frequency noise of the , iterative process is as follows:

[0143] (1)

[0144] (2)

[0145] (3)

[0146] in, is the filter coefficient, 0< <1, is the filtered position signal, The old spot position value information used in the iterative calculation, The calculation is to iterate the new spot position value information, and the iterative process is to repeat formula (1), (2), and (3) in sequence.

[0147] In step 3, the filtered spot position information needs to be corrected and fitted, and the fitting correction is completed in the drive control board 4 and used as the final spot position information.

[0148] The corrected fit is filtered The position signal is generated after the high-precision laser position sensor calibrates and corrects N positions. The method is to fix a metal cube with a receiving surface perpendicular to the X and Y axes on the lens, and the cube moves with the lens, such as Figure 3 As shown in the figure, when the voice coil motor of a certain axis drives the lens to move the cube, the actual displacement of the cube is measured by the laser position sensor placed parallel to the axis to correct and fit the spot position information solved by PSD. The entire travel range of a certain axis is divided into N equal parts, and the step length of each equal part is the total travel divided by N distance, that is, each travel moves in sequence according to the spot position value. , , … , among which The position is the zero position of the measurement starting point, and the high-precision laser position sensor values ​​of the corresponding N positions are collected at the same time , , … , after obtaining the two sets of numbers, the quadratic polynomial generated by mathematical tools such as Origin or MATLAB is as follows:

[0149]

[0150] in, is the spot position information after filtering on a certain axis, A, B and C are binomial fitting parameters, The spot position information finally generated after fitting is used as the position feedback of the axis. The role of fitting correction is to improve the accuracy of the spot position, thereby improving the accuracy of image motion compensation.

[0151] Due to factors such as installation errors and nonlinearity of PSD devices, the spot position after solution will deviate from the actual value. It needs to be calibrated by a high-precision position sensor, and then fitted and corrected to generate new spot position information as a feedback value. The fitting correction calculation is completed in the drive control board 4. The position accuracy of the position sensor used for fitting calibration is higher than that of the PSD. One implementation scheme uses the Omron ZW-S20 laser displacement sensor for calibration correction.

[0152] The operational amplifier required for the weak signal detection of the spot current in the PSD sensor signal processing board 3 is required to have low noise, high bandwidth and low bias current; the AD device selection requires high resolution, good linearity and the ability to synchronously sample the voltage converted at both ends of the one-dimensional PSD. One of the implementation amplifiers selected is the ADI LT1464A low bias current and high bandwidth instrumentation operational amplifier, and the AD device uses the ADI AD7606 with synchronous sampling of 16-bit resolution.

[0153] In one implementation scheme, the drive control board 4 MCU uses the GigaDevice ARM Cortex-M4 microcontroller GD32F450, which supports hardware floating-point operations and has a maximum main frequency of 200MHz.

[0154] The image motion compensation effect of the telephoto camera was tested under the disturbance condition of 2°@1Hz on a rocking platform. After the coarse compensation by the outer frame gyro, the residual image motion value was about 2 pixels. The residual image motion value was given as the input of the two-dimensional optical fast translational image motion compensation. After compensation, the root mean square value of the image motion at the exposure moment was controlled within 0.6 pixels, meeting the requirement that the image motion at the imaging exposure moment was no more than 1 pixel.

Claims

1. A two-dimensional optical fast translational image motion compensation method based on PSD, characterized in that: The method is based on a non-contact photoelectric displacement measurement device for translational image shift compensation, wherein the non-contact photoelectric displacement measurement device comprises a semiconductor laser module (1) and a one-dimensional PSD device (2); The number of one-dimensional PSD devices (2) is 2, which are x-axis and y-axis one-dimensional PSD devices (2), respectively, and are arranged in parallel to the x-axis and y-axis in the direction of lens movement, and do not move with the lens; The number of semiconductor laser modules (1) is two, which are used in conjunction with the one-dimensional PSD devices (2) of the x-axis and the y-axis, respectively, and move with the lens. When the device is stationary, the light beams emitted by the two semiconductor laser modules (1) are vertically irradiated to the center position of the photosensitive surface of the corresponding one-dimensional PSD device (2); The method specifically comprises: Step 1: The semiconductor laser module (1) emits a light beam, which is vertically irradiated onto the photosensitive surface of the one-dimensional PSD device (2) of the corresponding axis, and the light spot generates a photoelectric analog signal on the one-dimensional PSD device (2) of the axis; Step 2: Amplify the photoelectric analog signal and convert it into a digital signal; Step 3: According to the converted digital signal, position calculation, filtering and fitting correction are performed to obtain the final spot position information; Step 4: Integrate the residual value of the external frame gyro, convert the difference in angular velocity into an angle difference, convert the angle difference into a radian difference, and then multiply it by the focal length of the imaging system to obtain the linear displacement to be compensated. ; Step 5: Set the linear displacement of a certain axis As a given value, the spot position information after the axis is fitted As the feedback value, the feedback value is compared with the given value in real time. According to the positive and negative and size of the difference, the lens is controlled to make two-dimensional high-speed linear translation along the X and Y axes to compensate for the image shift. When the difference is zero, the image shift compensation is completed and the lens stops moving. In step 3, the position solution specifically includes: The spot position information of a one-dimensional PSD device (2) on a certain axis is obtained by the spot offset position relative to the center position of the PSD on the axis. The calculation formula is: in, and It is the AD code value of the photocurrent signal converted from the electrodes at both ends of the corresponding axis one-dimensional PSD device, and S is half of the effective length of the effective photosensitive surface of the one-dimensional PSD device. The calculated offset relative to the PSD center position of the axis is the spot position information; In step 3, filtering is performed on the light spot position information, and the filtering specifically includes: Use IIR low-pass filter to iteratively remove The high-frequency noise in the spot position information, the iterative process is as follows: （1） （2） （3） in, is the filter coefficient, 0< <1, is the filtered position signal, The old spot position value information used in the iterative calculation, The calculation is the new spot position value information used for iteration. The iteration process is to repeat formula (1), (2), and (3) in sequence. In step 3, the fitting correction is the filtered position signal The generated value is obtained after the high-precision laser position sensor calibrates and corrects N positions, specifically: A metal cube with a receiving surface perpendicular to the directions of the X and Y axes is fixed on the lens, and the cube moves together with the lens; When the voice coil motor of a certain axis drives the lens to move the cube, the actual displacement of the cube is measured by the laser position sensor placed parallel to the axis to correct and fit the spot position information solved by PSD. The entire travel range of a certain axis is divided into N equal parts, and the step length of each part is the total travel divided by N distance, that is, each travel moves in sequence according to the spot position value. , , … , among which The position is the zero position of the measurement starting point, and the high-precision laser position sensor values ​​of the corresponding N positions are collected at the same time , , … ,After obtaining two sets of numbers, the spot position information finally generated after fitting is obtained.

2. The method for compensating two-dimensional optical fast translational image motion based on PSD according to claim 1, characterized in that: In step 3, the fitting correction formula is: in, The final spot position information generated after fitting is used as position feedback of the control system. is the spot position information after filtering, and A, B, and C are binomial fitting parameters.

3. The method for compensating two-dimensional optical fast translational image motion based on PSD according to claim 1, characterized in that: In step 4, the linear displacement of a certain axis The calculation formula is: in, Give a given value of the inertial space angular velocity for a certain axis, in degrees / s. is the actual feedback angular velocity value of the gyro of this axis, in degrees / s. To control the control cycle of the system, is the focal length of the camera, and 57.3 is the angle radian conversion factor.

4. The method for compensating two-dimensional optical fast translational image motion based on PSD according to claim 1, characterized in that: The length of the photosensitive surface of the one-dimensional PSD device (2) is not less than the effective stroke of the lens on the corresponding axis.

5. A two-dimensional optical fast translational image motion compensation system based on PSD, characterized in that: The system is used to perform the method according to any one of claims 1 to 4, and the system comprises a non-contact photoelectric displacement measuring device, a PSD sensor signal processing board (3), a drive control board (4) and a push-pull voice coil motor (5); The contactless photoelectric displacement measuring device comprises a semiconductor laser module (1) and a one-dimensional PSD device (2); The number of one-dimensional PSD devices (2) is 2, which are divided into x-axis and y-axis one-dimensional PSD devices (2), which are arranged in parallel to the x-axis and y-axis in the direction of lens movement and do not move with the lens; The number of semiconductor laser modules (1) is two, which are used in conjunction with the one-dimensional PSD devices (2) of the x-axis and the y-axis, respectively, and move with the lens. The setting position is such that when the device is stationary, the light beams emitted by the two semiconductor laser modules (1) are vertically irradiated to the center position of the photosensitive surface of the corresponding one-dimensional PSD device (2), and the light spot generates a photoelectric analog signal on the PSD, and the signal can be used to calculate the position information of the light spot relative to the center position of the PSD photosensitive surface; There are two PSD sensor signal processing boards (3), one for the x-axis and the other for the y-axis, which are used to amplify the x-axis and y-axis light position analog signals, convert them into digital signals and send them to the drive control board (4); The drive control board (4) is used to perform light spot position calculation, filtering and fitting correction according to the digital signals after the x-axis and y-axis conversion, obtain the final light spot position signals of the x-axis and y-axis, and drive the push-pull voice coil motor (5) to drive the lens to perform two-dimensional linear motion; The number of push-pull voice coil motors (5) is 4, of which 2 are on the x-axis and 2 are on the y-axis, and they are arranged at both ends of the lens with the lens as the center, and are parallel to the x-axis and y-axis moving directions of the lens, respectively.

6. A computer device comprising a memory and a processor, wherein a computer program is stored in the memory, wherein: When the processor runs the computer program stored in the memory, the steps of the method according to any one of claims 1 to 4 are performed.

7. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a plurality of computer instructions, and the plurality of computer instructions are used to enable a computer to execute the method according to any one of claims 1 to 4.

Images