Manuscript received August 7, 2025; revised August 28, 2025; accepted November 11, 2025; published March 26, 2026.

Abstract—Automated analysis of astronomical images has become increasingly important due to the growing volume of data generated by modern sky surveys and observatories. This paper presents an image segmentation approach designed to improve the automated detection and classification of objects in astronomical frames. The approach combines classical image processing and modern computer vision techniques to isolate and characterize objects of interest, even in the presence of noise, telescope aberrations, low contrast, or overlapping sources commonly found in wide-field images. The segmentation pipeline employs adaptive thresholding, background subtraction, and morphological filtering to enhance the visibility of both point-like and extended sources. The classification process enables the differentiation between various imaging artifacts and celestial objects, including stars, galaxies, Small Solar System Objects (asteroids, comets), and even artificial satellites. The paper describes the modern features for astronomical image processing implemented in the Lemur software within the scope of the Collection Light Technology (CoLiTec) project (https://colitec.space). The Lemur software is designed to perform a sequence of the following main steps: data mining, classification, background alignment, brightness equalization, image segmentation, typical shape analysis, pattern recognition, object detection/identification, astrometric/photometric reduction, and trajectory detection. The Lemur software has facilitated over 1700 discoveries of asteroids, including 5 Near-Earth objects, 21 Trojan asteroids of Jupiter, and 1 Centaur. In total, it has been used in about 800,000 observations, during which five comets were discovered.

Keywords—image processing, filtering, segmentation, classification, pattern recognition, data mining, knowledge discovery, big data, astronomy

Cite: Sergii V. Khlamov, Vadym E. Savanevych, Vladimir P. Vlasenko, Oleksandr B. Briukhovetskyi, and Tetiana O. Trunova, "Astronomical Image Segmentation in Computer Vision for Automated Object Detection and Classification," Journal of Image and Graphics, Vol. 14, No. 2, pp. 163-171, 2026.

Copyright © 2026 by the authors. This is an open access article distributed under the Creative Commons Attribution License (CC-BY-4.0), which permits use, distribution and reproduction in any medium, provided that the article is properly cited, the use is non-commercial and no modifications or adaptations are made.