Research On Real-time Phase Detection Auto Focus Technique Of Long Focus Airborne Camera Based On R-C Optical System

With the continuous development of airborne remote sensing imaging technology and people’s demand for high quality aviation remote sensing image, the development of long focal length, high resolution aerial camera become the current urgent demand. But due to the changes of temperature and pressure of the camera from the ground to air environment, the focal plane of the camera will move and image quality would be affected. The size of the optical system is larger, so the impact is more serious for long-focus camera. Therefore, for ensuring the camera image quality, to take appropriate auto focus technology is one of the key technologies of aviation camera, especially long focal length aviation camera.Domestic and international aviation camera mainly adopts the auto focus methods of ground calibration method, optoelectronic autocollimation method, image auto focus method, etc. Based on the detailed analysis of the advantages and disadvantages of all kinds of auto focus methods, according to the camera structure characteristic and the outside environment, this paper proposes a real-time phase detection auto focus technique of the long focus airborne camera based on R-C optical system, and the relevant theories were studied.According to the requirements of the camera imaging, combined with ground calibration method, optoelectronic collimation detection method and the advantages of image auto focus method, this paper proposes a real-time phase detection based on image focal technology, solved the camera can’t real-time inspection focal problems in the process of taking pictures. Improved median filtering method is used for image preprocessing, 8 neighborhood using the improved Laplacian as focusing evaluation function, design a kind of difference of focal plane position opto-mechatronics integration structure, introduces in detail the inspection method related optical principle, structure composition, focal algorithm and the working process.This paper illustrates the sources of noise and the influence on image quality, analysis the influence of noise on focus. This article tells the methods of aviation image preprocessing. Through the analysis of the characteristics of median filtering, according to the different characteristics of the image edge points and noise points, an improved median filtering method is proposed. Through actual noise image validation, this method is effective to denoise and has a shrter processing time.The focusing evaluation function plays an important role. This paper introduces the common focusing evaluation function. Two groups of representative aerial images are selected to calculate and validate the performance of the focusing evaluation function. According to the analysis, improved 8 neighborhood Laplace algorithm is put forward, and its speediness and high sensitivity are verified. This paper analyses the motion blur effects on image quality, and illustrates the motion blur effect on focus.This article tells the necessity of focusing window selection, analyze the theoretical model of the focusing window model and aerial photographic depth of field. The central region and non-uniform extraction method is introduced. Based on the analysis of the aviation image characteristics, the golden region segmentation method for aerial images is proposed, and the mthod can reduce 72% of the amount of calculation. The feasibility of this method is verified by experiment.According to the characteristics of the phase detector focal technology, based on the ergodic search method and variable step climbing method, the suitable auto focus search algorithm is proposed, to ensure the speediness, the reliability of the auto focus search strategy and focus accuracy.Experiments prove that average auto focus accuracy of the real-time image auto focus method in phase detection auto focus technique is ±0.015 mm, and this accuracy can satisfy the demands of carbocoal deep accuracy. Aviation camera image at the low temperature low pressure environment and related data is obtained through thermal optical test, and the data verify the correctness of autocollimation focus method in the phase detection auto focus technique. The precision of the autocollimation focus method is ± 0.013 mm, its consumption time is only 26 s. Compared with the photoelectric autocollimation method based on grating, precision is double time, onsumption time is 1/3 of the original time. The experimental results show that the real-time phase detection auto focus technology can effectively compensate the influence of defocus.