depthanythingv2

Unofficial ComfyUI implementation of Video Depth Anything, for depth estimation on long videos without compromising quality, consistency, or generalization ability.

Merging YOLOv9 and DepthAnythingV2

[IEEE ICRA 2026] The official repository for paper WildCross: A Cross-Modal Large Scale Benchmark for Place Recognition and Metric Depth Estimation in Natural Environments at IEEE ICRA 2026

ArduMonoNav: MAV Navigation via Monocular Depth Estimation and Reconstruction

depth map generation using MiDaS, ZoeDepth, DepthAnything and Tiling

Real-time monocular ADAS perception pipeline using YOLOv8, Depth Anything V2, Kalman tracking, and BEV lane detection.

Real-time depth anything model 2 with GPU acceleration supports Apple silicon GPU's .

A state-of-the-art Assistive Spatial AI system that integrates Vision-Language Models (LLaVA v1.5) with Metric Depth Transformers (Depth Anything V2) to provide real-world spatial awareness.

Real-time autonomous navigation system powered by monocular depth estimation and object detection. Implements Depth Anything V2 for dense spatial awareness to enable intelligent obstacle avoidance without specialized LiDAR hardware.