A STUDY ON THE DEVELOPMENT OF OPEN SOURCE-BASED IMAGE MOSAICKING SOFTWARE

摘要

Image mosaicking can be handled by ENVI. Erdas Imagine, ArcGIS and other commercial software, but they are designed centering on the implementation of functions, resulting in producing very different outputs depending on the user's software handling skills. Thus, there is a need to develop automatic image mosaicking software based on the map sheet. This study presents the development of the automatic mosaicking and color correction software for the national land satellite information under the project of establishment and operation of the satellite information utilization center, and also conducts the development of color correction algorithm. The color correction used the histogram matching method. In order to check the image stretching state of the NIR bands of N number of images in a different color distribution, the standard deviation of each NIR band was calculated, and the NIR band with the biggest standard deviation was made to become the reference for selection. The reference NIR was used to correct the color between the other NIR bands, and a merged NIR band was created through image mosaicking. The same method was applied to the color correction of the remaining R, G. and B bands. The color correction algorithm was developed based on the open source GDA1. library using Python. This study targeted the low-capacity images, and the function is implemented manually. In order to boost the utilization of satellite images in the development of a color correction algorithm, the accuracy should be verified and the speed improvement and other relevant algorithms should be improved according to standard deviations, targeting the N number of large-capacity images so as to automate the function.