Research On Infrared Remote Sensing Image Mosaic Technology

Whether in the military field or in the relevant experimental research in the civilian field,the demand for larger field of view and higher resolution images is becoming more and more urgent.In terms of detectors,infrared remote sensing imaging is one that can be used 24 hours a day.Uninterrupted hours of operation,while avoiding the efficient detection of common electromagnetic interference.Infrared remote sensing image splicing technology developed in combination with market demand and the advantages of infrared remote sensing detectors is increasingly used in the military field,which plays an important role in China's national defense construction.Infrared remote sensing image stitching is a major requirement in practical applications.First,the image has a relatively high registration accuracy,and second,a higher algorithm running speed is required.Based on the research methods adopted by previous research institutes,this paper will optimize the existing deficiencies in infrared remote sensing image research methods,aiming to develop an effective real-time splicing of infrared remote sensing images.The method enables the near-real-time high-resolution,large-field infrared remote sensing seamless mosaic image to be obtained.Based on this purpose,this paper studies and improves the ORB feature extraction operator,and introduces the DAISY fast algorithm to describe the feature points,and introduces a high-dimensional binary search tree innovatively,and proposes a new infrared.Remote splicing technology for rapid splicing.This paper analyzes the imaging characteristics of infrared remote sensing images,expounds the imaging mechanism of infrared remote sensing images and the realization principle of image stitching technology,and discusses the following issues in the process of infrared remote sensing image stitching:(1)There are usually some extreme or minimum values on the infrared remote sensing image,which is caused by the salt and pepper noise existing in the infrared remote sensing system;after the specific statistical analysis of the histogram of the infrared remote sensing image,the gray is found in the image.The difference between the maximum value and the minimum value of the degree is small,that is,the gray value of the image as a whole tends to be concentrated in a small range,which means that the image contrast is low.Therefore,in order to reduce the influence of salt andpepper noise on the imaging quality of infrared remote sensing images,this paper uses the anisotropic filtering method with better filtering effect to filter the salt and pepper noise of the image,and then uses the dual platform histogram equalization method to infrared remote sensing image.The grayscale range has been expanded to greatly increase the contrast of the image.(2)In order to improve the splicing speed of infrared remote sensing images,this paper makes an innovation and optimization for the feature point extraction operator ORB.First,the image scale pyramid is constructed,feature points are extracted at different scales,and the gray scale centroid method is introduced.Compared with the traditional ORB algorithm,the optimized feature point extraction operator can ensure that the registration accuracy of the algorithm is not affected when the image is scaled and rotated.Secondly,in order to improve the accuracy of infrared remote sensing image matching,the high-dimensional binary search tree is introduced to match the feature points,and the calculation time and matching accuracy of the algorithm are improved.Finally,the RANSAC matrix is used to reduce the mismatched feature point pairs.(3)In order to eliminate the gap after the infrared remote sensing image is stitched and ensure the running speed of the algorithm,an improved fade-in and fade-out weighting algorithm is proposed to fuse the image.Traditional image fusion algorithms often rely on empirical values when determining weighting factors.In this paper,the weighting factor is determined according to the distance relationship between the feature points and the two images,without human intervention,so that the stitched image can achieve a smooth transition.After the image splicing is completed,the spliced image is subjectively judged by visual evaluation,and then the objective parameters are analyzed by different index parameters.Finally,the two parts are combined to obtain the reasonable evaluation of the image quality of the infrared remote sensing image after fusion.Compared with the traditional image stitching method,the experimental flow designed in this paper can improve the running speed of the fusion algorithm and achieve high-speed and high-quality stitching of images based on the quality of infrared remote sensing image stitching.The experimental results show that the image matching accuracy of the proposed algorithm is improved by 4.1% compared with the ORB algorithm,and the running time of the algorithm is 63.8 times higher than that of the SITF image stitching algorithm.The optimized infrared remote sensing imagestitching algorithm runs stably and has good real-time and robustness.