Analysis of radiometric calibration on matrix imager of LAPAN-A3 satellite payload

In common satellite image preprocessing, quality of radiometric and geometric correction play important roles in producing a high quality image. Both radiometric and geometric correction needs accurate calibrated camera parameters. This paper presents the analysis of radiometric calibration that has been done on matrix imager of LAPAN-A3 satellite payload. In this research, radiometric process of LAPAN-A3 matrix imager is modeled using a linear model, and then the model is identified using linear least-square regression method. Series of calibration measurements are taken using integrating sphere as a highly uniform light source. Experiments results show that for relative radiometric correction, common linear regression method has a good overall performance but for absolute radiometric correction purpose, the error on low-radiance of band-3 is quite high. On the other hand, conventional method has a good performance on low-radiance area but has significantly lower accuracy overall. By combining both calibration methods, radiometric correction produces a very good balance between low-radiance quality and overall accuracy. These results should be used as references for further development of LAPAN-A3 image correction algorithm, particularly on radiometric correction part.