Research On Image Real Time Definition Improvement For Space Optical Remote Sensing Images

In recent decades,China's space remote sensing technology has been developed rapidly,and remote sensing images are widely used in military and civilian fields.Due to degradation factors such as atmospheric disturbance and electronic signal conversion in the imaging link of optical remote sensing images,the degradation of remote sensing images is caused and the sharpness of remote sensing images decreases,which seriously affects the application value of remote sensing images.In order to improve the resolution of remote sensing image and recognition,we summarized and analyzed the influence factors of resulting in a decline in image resolution.In order to clarify the remote sensing images disturbed by thin clouds and mists in the atmosphere and the low illumination images generated by insufficient illumination conditions and clouds occlusion,we have completed the following five main tasks:(1)The characteristics of space remote sensing technology and the development of space remote sensing satellites at home and abroad were introduced in this paper,we discussed the main factors affecting the imaging performance of cameras,and analysed the importance on improving clarity of remote sensing images,then the research background and significance were introduced,what's more,we summarized the research status of cloud removal methods of remote sensing images and enhancement methods of low-illumination remote sensing images at home and abroad.(2)The system composition and imaging principle of the space remote sensing camera were introduced.The imaging mechanism of the cloud image is analyzed in detail,which provides a theoretical basis for the research of the clear processing method of the thin cloud image.The characteristics of low illumination remote sensing image were analyzed.The evaluation indexes of remote sensing image clarity were studied from two aspects of reference image evaluation and non-reference image evaluation,which could be used for the evaluation of the effect of clarification processing algorithm.(3)A defogging method for single remote sensing image based on multivariate linear regression model was proposed.Several image characteristic parameters related to fog concentration were analyzed.A multivariate linear regression model was established.The parameters in the model were learned by gradient descent algorithm,so that the effective estimation of transmittance could be obtained.Then,the atmospheric ambient light could be obtained by weighted quadtree method.Finally,the haze-free images were restored according to the cloud and fog image degradation model.The experimental results show that these methods achieved good results in the removal of thin cloud and fog.A cloud and fog removal method for single remote sensing image in HSI color space was proposed,which changeed the image into HSI color space for processing,and the color information in the original image was protected to a certain extent.Compared with the single image defogging method based on feature learning model proposed in previous,the experimental results show that the image processed by this method had better visual effect and better color fidelity.(4)The enhancement methods for low illumination remote sensing images were studied.An improved multi-scale Retinex algorithm combined with local contrast adaptive adjustment was proposed to improve image quality.The logarithmic function of multi-scale Retinex algorithm was replaced by Sigmoid function to reduce data loss.In order to improve local detail information,the improved multi-scale Retinex algorithm was used.The proposed algorithm improved the local contrast of the image by the adaptive local contrast enhancement method,and effectively improved the visual effect of the low il umination remote sensing image.(5)In order to verify the performance of the remote sensing image sharpening processing algorithm in this paper,the processing experiment based on image hardware processing platform was carried out for the remote sensing image sharpening processing.The experimental results show that the algorithms can effectively enhance the image sharpness and improve the visual effect.The algorithms can basically meet the real-time requirements and be as reference for the practical application.In this paper,the real-time sharpness processing technology of space optical remote sensing image was deeply studied.The methods of sharpness processing of thin cloud and mist interference remote sensing image and low illumination remote sensing image were proposed.The related experimental analysis and verification were carried out.The results show that the sharpness of remote sensing image has been improved,and our results have a good performance in visual effect.These methds could be implemented on the image hardware processing platform,and basically meet the real-time requirements.