Research And Application Of Optical Image Registration Algorithm In Infrared And Panchromatic Bands

Since the concept of "digital earth" was put forward,remote sensing image data containing geographic information has played a vital role in achieving this goal.At the same time,geographic information data has played an important role in the country's infrastructure construction and national defense engineering.The field of environmental monitoring plays an irreplaceable role.Among them,the digital orthophoto(DOM)produced according to the acquired original image is one of the most important 4D surveying and mapping products.It has high position accuracy and can directly see the characteristics of the image and can be used as control information.Evaluate other data to obtain the quality evaluation results of other data,which can be used to update basic map data,draw thematic maps,or directly construct digital threedimensional cities on the basis of DOM.At the same time,it can also extract various natural and humanistic information needed by itself from the DOM to facilitate subsequent urban planning consultation and planning design.DOM is widely used in the production and life of all walks of life.The quality of DOM is related to the accuracy of subsequent applications and analysis.Higher and higher requirements are put forward for the position accuracy and image quality of DOM.Image quality assessment has strong practical significance.In view of the above problems,this paper studies the matching algorithm of infrared images and panchromatic images,overcomes the problem of different image gray values due to different radiation between different bands,realizes the matching between multi-source data,and design a software is used for automatic quality evaluation of DOM.The above algorithm is used to calculate the position accuracy of the DOM,and the results are analyzed,and it is detected that the mosaics in the DOM are misaligned,areas with clouds,and tensile deformations areas.At the same time,a series of objective evaluation indicators are used to evaluate the quality of the image itself,and an objective quality evaluation document is generated.The main research content and results of this thesis are as follows:(1)Propose an infrared image matching method combining mutual information and multiple features.By building an image pyramid,matching layer by layer to obtain the relative relationship between the images,in each layer by using the edge features of the image,the useful information of the image is preserved while simplifying the calculation of mutual information,and finally RANSAC is used to eliminate the wrong matching point pairs.,Get the correct matching result.Through comparison and analysis of different methods,the method in this paper accelerates the speed of image matching,improves the accuracy of image matching,and has higher reliability.It is an excellent matching method for infrared and optical images.(2)Design and produce a DOM quality assessment software.The DOM quality automatic evaluation software provides users with a visual interface display and an interactive data selection program,which realizes the reading and writing of files,image reading and writing,image display,image quality evaluation,and memory cleaning(continuous processing))And other functions.By calculating the accuracy evaluation indicators such as the position accuracy of the DOM image,the edge accuracy,and the image quality evaluation indicators such as brightness,sharpness,and contrast,based on the above indicators,an objective evaluation conclusion can be drawn for the image to be evaluated.At the same time,there is a certain solution to the problem of large stitching misalignment that may exist in large DOM images,and the location of this phenomenon is reflected in the final image quality evaluation file.