Result listed — see original announcement
Original award wording
Finalists
Project profile
CARN
An autonomous search-and-rescue operations platform centered on drone video, intended for rescue teams and drone operators. Its software portal and mission workflow do not establish that this project built drone hardware.
The submission describes fine-tuned YOLOv8 processing visible-light and thermal footage to produce detections and tracks, with GPS alerts for rescuers. A portal registers missing people and manages missions and waypoints. Three Claude agents coordinated through LangGraph handle planning, detection review, briefings, and operations.
The creator presents it as augmenting human rescue judgment rather than replacing it. Speed, training-data size, alert thresholds, and effectiveness are creator claims; the platform has not been run or flight-tested.
Creator or team (original public names)
Kevin Collins; the page heading reads Echofold
Links & sources
Links may have changed since we checked them. Descriptions from creators have not been independently tested.
Details in the original language
Original title
CARN
Original creator credit
Kevin Collins；页面heading为Echofold
作品是什么
围绕无人机视频的自主搜救运营平台；有软件门户和任务协作说明，但不足以证明本作制作了实体无人机硬件。
给谁用
搜救团队与无人机操作员。
问题或体验
从航拍视频发现失踪人员并及时把GPS位置送给救援队，协调多架无人机的任务及检测信息。
如何使用
作者说明以两轮微调的 YOLOv8 处理可见光与热成像输入，输出置信度和轨迹；门户登记失踪人员、安排任务和航点，分级告警；LangGraph 中三个 Claude agents 进行任务规划、检测复核、简报和运营协调。
得到什么
门户、实时检测及GPS告警、任务和多机协作。作者强调扩展搜救团队能力而非代替判断；速度、训练集规模和告警阈值属于作者声明，不是本轮测量。
关键机制
YOLOv8 检测可见光与热成像视频；LangGraph 中三个 Claude agents 协调任务和复核。均是投稿说明，未实飞。
特点
把视频检测、GPS 告警和搜救任务协作连接起来；作者将它定位为辅助搜救人员判断，非替代人工。
Events & award announcements
Read original announcement
Sources last checked: 2026-10-11