Max Robotics — Global Service Robot Directory + US Certification

About

Zhiyu Xiang is a computer vision and robotics researcher whose work spans over two decades, encompassing robot localization, visual SLAM, 3D reconstruction, and autonomous driving perception. His research career began with foundational contributions to mobile robot navigation, including laser-based indoor localization and map building techniques developed in the early 2000s. Over time, his focus evolved toward increasingly sophisticated visual sensing systems, culminating in influential work on depth completion and scene understanding for autonomous vehicles. Xiang's most cited contribution, "DenseLiDAR" (2021, 57 citations), introduced a real-time pseudo dense depth guided network that significantly advances depth completion from sparse LiDAR inputs — a critical challenge in autonomous driving. His earlier work on binocular visual odometry (2013, 27 citations) demonstrated high-accuracy, high-frequency localization using GPU-accelerated bundle adjustment. He has also tackled multi-task learning, jointly addressing semantic segmentation and depth completion with boundary constraints (2020, 20 citations), and developed robust SLAM systems capable of handling weakly textured environments. His research into catadioptric multi-stereo systems and learning-based stereo matching further underscores his breadth across geometric and deep learning approaches to 3D perception, making his work highly relevant to robotics and autonomous systems communities alike.