TSolar images from space telescopes contain a wealth of information on solar variability, of great importance both in solar physics and in view of Space Weather applications. Obtaining this information, however, requires the ability to process large amounts of data over long periods in an objective fashion. In previous work, we have proposed a multi-channel unsupervised spatially-constrained multichannel fuzzy clustering algorithm (SPoCA) that automatically segments EUV solar images into Active Regions (AR), Coronal Holes (CH), and Quiet Sun (QS). Applying SPoCA to SoHO-EIT images on almost the full 23rd solar cycle, we obtained variations of area, mean intensity, and relative contributions of AR, CH, and QS to the solar irradiance, consistent with previous results. The Royal Observatory of Belgium is a co-investigator on the SDO Science Center, a suite of software pipeline modules for automated feature recognition and analysis of the Solar Dynamics Observatory data. As such, we will deliver our Active Region segmentation tool, SPoCA, to the SDO Science Center, where it will be inserted into the SDO pipeline at Lockheed Martin Solar and Astrophysical Laboratory to run in near real time on SDO-AIA data. In the present poster, we present the fine-tuning of the algorithm and its implementation for optimal segmentation and performance. We show how to combine SPoCA’s detection of AR on subsequent images in order to allow for automated tracking and naming of any region of interest, paving the way for systematic temporal follow-up studies of AR, CH, and QS. Finally, we indicate how to tailor the method to the needs of the high resolution EUV images soon to be delivered by the EUV telescope on board the Solar Dynamics Observatory.