feat(perception): Semantic scene understanding — room/environment classification + object detection #141

Closed
opened 2026-03-02 09:31:27 -05:00 by sl-jetson · 0 comments
Collaborator

Summary

Scene understanding for context-aware robot behavior.

Requirements

  • Room classification: Detect room type (kitchen, living room, hallway, outdoor, garage) from visual features
  • Object detection: YOLOv8-nano TensorRT for common objects (chairs, tables, doors, stairs, pets)
  • Hazard detection: Identify stairs, drops, wet floors, glass doors
  • Context topics: /social/scene/room_type, /social/scene/objects (DetectionArray)
  • Behavior adaptation: Robot adjusts speed/personality based on scene context
  • Nav2 integration: Objects as dynamic obstacles in costmap
  • Performance: 15+ FPS on Orin with YOLOv8-nano TRT FP16

Agent: sl-perception

Labels: social-bot, perception

## Summary Scene understanding for context-aware robot behavior. ## Requirements - **Room classification**: Detect room type (kitchen, living room, hallway, outdoor, garage) from visual features - **Object detection**: YOLOv8-nano TensorRT for common objects (chairs, tables, doors, stairs, pets) - **Hazard detection**: Identify stairs, drops, wet floors, glass doors - **Context topics**: /social/scene/room_type, /social/scene/objects (DetectionArray) - **Behavior adaptation**: Robot adjusts speed/personality based on scene context - **Nav2 integration**: Objects as dynamic obstacles in costmap - **Performance**: 15+ FPS on Orin with YOLOv8-nano TRT FP16 ## Agent: sl-perception ## Labels: social-bot, perception
Sign in to join this conversation.
No Label
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: seb/saltylab-firmware#141
No description provided.