Max Robotics — Global Service Robot Directory + US Certification

About

Dimitrios Katsoulas is a computer vision researcher whose work focuses on robotic perception and 3D object localization, particularly for industrial automation. His key research areas include range image analysis, superquadric modeling, and vision-guided robotic manipulation. Katsoulas made significant contributions to the depalletizing problem—the automated unloading of stacked boxes—by developing systems that use time-of-flight laser sensors and 2.5D range imagery to enable real-time, lighting-independent object detection. His most cited paper, "Superquadric Segmentation in Range Images via Fusion of Region and Boundary Information" (2008, 31 citations), advanced the use of superquadrics as modeling elements for multiple object segmentation, a technique with high potential in computer vision. His earlier work on depalletizing systems (2002, 25 citations; 2003, 18 citations) demonstrated practical, robust solutions for industrial robotics, including edge detection and line fitting to extract 3D vertices for box localization. Katsoulas also contributed to the recovery of piled deformable objects and box-like superquadrics, addressing the challenging robotic bin-picking problem. His research, spanning from 2002 to 2008, has accumulated over 90 citations, reflecting its impact on automated vision systems for manufacturing and logistics.