The Paper Presents a deconvolution method based on the double passthrough the human eye By using an optical-digital-computer. S system, the methodsolves the point-spread function and the optical transfer function of t...The Paper Presents a deconvolution method based on the double passthrough the human eye By using an optical-digital-computer. S system, the methodsolves the point-spread function and the optical transfer function of the optical systemof eye from the reflective image of retina, reconstructs the retinal image.As the double pass feature says, in the Positive direction and negative direction ofthe light propagation, the amplitude spread function is the same along the same media ofeye. If He-Ne laser beam is used as a short-time point impulse source, the incident lightcan be referred to as coherent light. Because the surface of retina is not fine smooth,usually the reflection is in company with the scattering, the reflective light beam is considered as incoherent light. After many times of flashings, samplings and averagings,the reflective ratio of retina will approximate to a constant and be neglected. Therefore,the conclusion is: the light intensity distribution outside of eye is the auto-convolution ofthe light intensity distribution of the renal image, i. e. the auto-convolution of thePOint-spread function. By the convolution theorem in the spatial frequency domain,two-dimensional Fourier transform of the point-spread runction, i. e. the optical transfer runction (OTF), can be calculated. Sequentially, by inverse Fourier transform, thePOint-spead function (PSF) is obtained. Thus the de-auto-convolution is accomplished.If a real picture replaces the point light source, the picture image on the fovea near thelight axis and then is reflected incoherently, the intensity distribution of the receivedimage outside of eye is equal to the crossconvolution between the intensity distribution ofretina and the point-spread function. By doing deconvolution once more. i. e. once decross-convolution, the retinal image of the picture can be reconstructed. The results ofthe experiments show that the hybrid system is advanced at the objectivity, auto-calibration and dynamic recording.展开更多
文摘The Paper Presents a deconvolution method based on the double passthrough the human eye By using an optical-digital-computer. S system, the methodsolves the point-spread function and the optical transfer function of the optical systemof eye from the reflective image of retina, reconstructs the retinal image.As the double pass feature says, in the Positive direction and negative direction ofthe light propagation, the amplitude spread function is the same along the same media ofeye. If He-Ne laser beam is used as a short-time point impulse source, the incident lightcan be referred to as coherent light. Because the surface of retina is not fine smooth,usually the reflection is in company with the scattering, the reflective light beam is considered as incoherent light. After many times of flashings, samplings and averagings,the reflective ratio of retina will approximate to a constant and be neglected. Therefore,the conclusion is: the light intensity distribution outside of eye is the auto-convolution ofthe light intensity distribution of the renal image, i. e. the auto-convolution of thePOint-spread function. By the convolution theorem in the spatial frequency domain,two-dimensional Fourier transform of the point-spread runction, i. e. the optical transfer runction (OTF), can be calculated. Sequentially, by inverse Fourier transform, thePOint-spead function (PSF) is obtained. Thus the de-auto-convolution is accomplished.If a real picture replaces the point light source, the picture image on the fovea near thelight axis and then is reflected incoherently, the intensity distribution of the receivedimage outside of eye is equal to the crossconvolution between the intensity distribution ofretina and the point-spread function. By doing deconvolution once more. i. e. once decross-convolution, the retinal image of the picture can be reconstructed. The results ofthe experiments show that the hybrid system is advanced at the objectivity, auto-calibration and dynamic recording.
