| With the development of remote sensing technology,domestic high resolution remote sensing image resolution has already reached sub-meter level,widely used in agriculture,forestry,geological survey,urban planning and other fields,the quality of remote sensing image requirements are becoming higher and higher,corresponding to the emergence of the evaluation of image quality of various methods.However,for mine remote sensing and other fields that have high requirements for local image quality but relatively few requirements for overall image quality,traditional image quality evaluation methods may not meet the requirements.There may be a situation that the overall quality of the image meets the requirements,but the quality of the local area of concern of the image is not qualified.Therefore,it is particularly important to find a method to evaluate the local quality of images.In this paper,the quality of images in the mining area is evaluated mainly through the evaluation of the radiation accuracy and the geometric correction accuracy,and the quality of images in the mining area is improved according to the evaluation results.The main research contents and conclusions are as follows:(1)An evaluation method for radiometric correction accuracy of remote sensing images based on Pixel Block segmentation is proposed,which improves the traditional evaluation method based on the whole image,and takes the image segmentation as the basic unit of evaluation.Then,each unit calculates the test factors of selected radiometric correction accuracy respectively,and finally carries out comprehensive evaluation calculation.The evaluation results of radiometric correction accuracy of image details are obtained.Results show that the remote sensing image radiation correction based on Pixel Block segmentation accuracy evaluation method is a good way to response the image detail feature and the quality of the mining area of mainland radiation correction,the correction precision of image regions of radiation in the results by different color intuitive said,for the assessment of the quality of image detail and mine the internal image to provide the reference.(2)Put forward a method to improve mining area the precision of geometric correction,geometric correction accuracy within the mining area is affected by various factors tend to be low,in this paper,through calculating the image geometric correction error global moran and local moran index,calculated the image the distribution of the geometric correction of errors and finding the image geometric correction error is bigger and stronger regional distribution correlation,Then,according to the influence of terrain on geometric correction,the number and distribution of geometric correction control points in areas with low geometric correction accuracy are adjusted,and geometric precision correction is carried out,so as to achieve the purpose of improving the geometric correction accuracy in mining areas without affecting the geometric correction accuracy of the whole image.The traditional geometric correction method is improved in order to improve the accuracy of image mining area.(3)A feature tonal consistency method by judging the image mining area,through all the same mine feature of statistical test area grayscale average,calculation error in the grayscale average of all the similar features,and then based on the grayscale average error is set limit,through feature in the grayscale average double error and error respectively in two times to complete inspection for mine terrain,It provides guidance for the selection and use of mine features. |
