CN202535462U - Rapid reflector image stabilization system based on image correlated detection - Google Patents

Abstract

The utility model discloses a rapid reflector image stabilization system based on image correlated detection, wherein a light beam is deflected by a rapid reflector to stably track the jittering optical axis of a photoelectric imaging system in a motion or vibration environment. Firstly, the deviation amount of the unstable optical axis is detected in real time in an image correlated manner and the deviation amount of the optical axis is fed back to the controller of the rapid reflector. Then the deflection of the rapid reflector is controlled in real time so as to stabilize the optical axis. The system can stabilize images within a single frame and the image blurring phenomenon of a high-resolution camera due to the jittering of the optical axis within integral time is eliminated. The imaging quality of an imaging system is improved.

Description

A kind of fast mirror based on the image coherent detection surely looks like system

Technical field:

This patent relates to the steady picture technology of photo electric imaging system, is specifically related to a kind of fast mirror based on the image coherent detection and surely looks like system.

Background technology:

Airborne, carrier-borne and aerial image system is during to target imaging; Can be because motion, flutter and the attitude variations etc. of platform make camera when exposure, have relative motion between photographed object image and photo-sensitive cell; Bring image blur and smearing with this, make image quality often produce degeneration to a certain degree, contrast and resolution reduce; For the high-resolution camera, situation is all the more so.The perhaps simple later image processing of resolution that improves image device is depended merely in this degeneration can not satisfy the requirement that obtains the high-resolution picture element; Therefore, how to reduce even eliminate external interference also becomes steady picture technology to the influence of image quality key point.

At present commonly used surely mainly contain as technology: optical profile type surely as technology, micromechanics formula surely as technology, electronic steady image is technological and comprehensive steady picture is technological.Optics is some optical elements to be set as the compensation to unstable image in light path as technology surely.Typical optics image stabilization method is to utilize wedge to control the direction of sight line, through moving or rotate wedge, changes the angle and direction of emergent ray, carries out the compensation of image drift, thereby reaches the purpose of steady picture.Optics is to be only applicable to vibrate less environmental condition as the major defect of technology surely, and the structure of compensator and manufacturing process are too complicated.Simultaneously, if only rely on optical elements such as prism, speculum or wedge to carry out passive compensation, stabilizing power receives than limitations.The electronic steady image technology is that the Applied Digital image process method comes directly to confirm the skew of image sequence and the technology that compensates; The electronic steady image method is just removed sequence of video images interframe and is not intended to the disturbance that shake causes because of video camera; It can only make video sequence in playing process, visually produce a kind of effect of stable playback; And can not in the time for exposure, carry out surely picture to single image, just can not really improve the resolution of its imaging.Micromechanics is to move through rotation or imaging focal plane that micro-mechanical device is directly controlled deflecting mirror as technology surely, and steady picture task is accomplished in the compensation image drift.Micromechanics surely looks like simple, the compact conformation of system, but need obtain position of focal plane and the capable control of contraposition shift-in through certain method.Be exactly to use the steady picture of the multiple common realization of steady picture technology surely comprehensively, to obtain the technology of better picture element as technology.

This patent is described surely to be belonged to the micromechanics formula as system and surely looks like system; Micromechanics image stabilization formula is the major tuneup to traditional ray machine image stabilization formula; Have high speed, high-precision image drift detectivity and compensation executive capability; And it is little, low in energy consumption, little and to numerous advantages such as the picture element influence are little to the influence of remote sensor optical system that the system that utilizes this technology to form also has a volume, so the micromechanics compensation way becomes one of first-selected steady picture technology of high-definition remote sensing device.

Summary of the invention:

Be to solve the deficiency that exists in the prior art, this patent provide a kind of fast mirror surely to look like system based on the image coherent detection.

The structure of the system of this patent is as shown in Figure 1, and whole system mainly is made up of imaging optical system 1, fast mirror 2, big face battle array focus planardetector 3, two miniature high-speed cmos detectors 4 and FPGA image relevant treatment unit 5.(2 adopt piezoelectric ceramic to put mirror soon to wherein said fast mirror; Described two miniature high-speed cmos detectors (4) are installed in big face battle array focus planardetector (3) both sides side by side, are no more than 2mm with big face battle array focus planardetector distance;

Optical signalling through imaging optical system through the reflection of fast mirror (2) after with signal gathering at big face battle array focus planardetector (3) with above two auxiliary miniature high-speed cmos detectors (4); Two auxiliary miniature high-speed cmos detectors (4) carry out high-speed sampling to the image of motion; Obtain the side-play amount of image through the relevant method of realtime graphic in the focal plane; Provide the side-play amount of image in real time through FPGA image relevant treatment unit (5); Side-play amount feeds back to the controller of fast mirror, and the deflection of control fast mirror makes image stable on the focal plane; Last big face battle array focus planardetector carries out integration to stable image, and so big face battle array focus planardetector has just obtained high-resolution image.

Image related algorithm amount of calculation is bigger, but precision of calculation results is very high, and error is less than 0.3 pixel.For the bigger problem of image correlation computations amount, this system design based on the rapid image correlation module of FPGA, FPGA has parallel processing, the characteristics of flexible programming.FPGA image related process module in this system adopts the method for fast Fourier transform to realize the real-time related operation of two two field pictures, and its inside function module is as shown in Figure 3, and its flow process of carrying out relevant treatment is following:

1. be sent to the image input interface of FPGA by the image of the miniature cmos detector of high speed output, at first pass through the image pretreatment unit of FPGA, image is carried out the pretreatment operation of related algorithm, i.e. undercurrent correction, gain calibration etc.

2. after input picture carries out preliminary treatment, to reference frame image and the parallel FFT Fourier transform of carrying out of current frame image, adopt base-2 fast fourier transform algorithms here respectively, improve arithmetic speed.

3. after reference frame image and current frame image carry out the FFT conversion; FFT to reference frame image gets conjugation again; With the direct multiplication mutually of the FFT of current frame image, with take advantage of the result carry out the IFFT inverse Fourier transform again, so just accomplished the related operation of two two field pictures.

4. at last through surface fitting, get peaked position among the related operation result, position coordinates is exactly the position offset between two two field pictures, gives fast mirror drive controlling input through the FPGA output interface.Accomplish the FEEDBACK CONTROL of side-play amount.

This system realizes that through the relevant method of high speed FPGA image the motion excursion amount of image detects; Adopt the rotation control chart picture of micro-mechanical device fast mirror stable on the focal plane then, the system that makes has advantages such as response speed is fast, volume is little, compact conformation.

Description of drawings:

The fast mirror that Fig. 1 is based on the image coherent detection surely looks like the system configuration sketch map;

Among the figure:

1---imaging optical system;

2---fast mirror;

3---big face battle array focus planardetector;

4---the miniature high-speed cmos detector;

5---FPGA image relevant treatment unit.

Fig. 2 is an imaging system focus planardetector scheme of installation;

Among the figure:

6---the imaging system focal plane.

Fig. 3 is a FPGA image relevant treatment Elementary Function block diagram.

Embodiment:

Surely look like system according to the described fast mirror of specification based on the image coherent detection; The structural representation that its is realized is as shown in Figure 1, and platform is made up of imaging optical system 1, fast mirror 2, big face battle array focus planardetector 3, two miniature high-speed cmos detectors 4 and FPGA image relevant treatment unit 5.

Imaging optical system 1 adopts common optical telescope, and bore is 150mm, and focal length 1800mm, F number are 12, resolution＜5 ", instantaneous field of view of system is 8urad, 0.95 ° of total visual field;

Fast mirror 2 adopts four fulcrum XY axial compression electric deflection platforms, and its closed loop angle of inclination can reach+2mrad, and resolution reaches 0.05urad, and the minute surface diameter is 50mm, the closed loop linearity 0.2%, and resonance frequency 3.3KHz, the distance H of its minute surface and focal plane is 100mm;

Big face battle array focus planardetector 3 adopts the monochromatic area array CCD device with global shutter, and its face battle array size is 1k*1k, and response wave length is 400～1000nm, and pixel dimension is 14um*14um, frame frequency 5fps, and can adjust the time of integration;

Miniature high-speed cmos detector 4 adopts high frame frequency facet battle array cmos device, and its face battle array size is 64*64, visible light wave range, and image drift is of a size of 8um*8um, frame frequency 150fps, the time of integration is less than 1ms; FPGA image relevant treatment unit 5 is as shown in Figure 3, and it is made up of image pretreatment module, image FFT conversion module, matrix IFFT inverse transform module and surface fitting module.The fpga logic unit number is 12060LEs, built-in 2 PLL and 234Kbits RAM, and external clock is 40MHz.

Claims (1)

1. the fast mirror based on the image coherent detection surely looks like system; It is made up of imaging optical system (1), fast mirror (2), big face battle array focus planardetector (3), two miniature high-speed cmos detectors (4) and FPGA image relevant treatment unit (5), it is characterized in that:

Described fast mirror (2) adopts piezoelectric ceramic to put mirror soon;

Described two miniature high-speed cmos detectors (4) are installed in big face battle array focus planardetector (3) both sides side by side, are no more than 2mm with big face battle array focus planardetector distance;

Optical signalling through imaging optical system through the reflection of fast mirror (2) after with signal gathering at big face battle array focus planardetector (3) with above two auxiliary miniature high-speed cmos detectors (4); Two auxiliary miniature high-speed cmos detectors (4) carry out high-speed sampling to the image of motion; Obtain the side-play amount of image through the relevant method of realtime graphic in the focal plane; Provide the side-play amount of image in real time through FPGA image relevant treatment unit (5); Side-play amount feeds back to the controller of fast mirror, and the deflection of control fast mirror makes image stable on the focal plane; Last big face battle array focus planardetector carries out integration to stable image, and so big face battle array focus planardetector has just obtained high-resolution image.

Images