Digital image processing is an important step before next processing, e.g. computer vision, prediction, and other image manipulations. Image processing is an input for another computational task. In the first time, image processing is learned by computer scientists with some mathematical processing, e.g. edge detection, noise filter, etc. but to day some discipline, e.g. remote sensing, GIS, use this subject as the main and important material.

I have a lot of friends from computer science that try to focus on this subject. He/she asked about the current trend of this interesting subject. Image processing now has been established, with some images to be processed should have a lot of band frequencies captured by satellite, e.g. Landsat, Sentinel, IKONOS, etc, called multispectral and hyperspectral (more than 13 band frequencies) data. So, if we only have one band frequency, it is difficult to manipulate this single band image (called panchromatic). But for computer scientist, it is interesting to focus on computer vision, whether video or object detection. But some areas, e.g. remote sensing, medical imaging, to say only a few, still need deep learning-based image processing. I’m surprise when see the curriculum in my previous campus that the digital image processing is held by remote sensing & geographic information systems (RS-GIS) department.

If you want to see how image processing for land use/cover classification, my Youtube video might be useful, especially for environmentalists.