Communitygithub.com

curiositech/windags-skills

3D電腦視覺標註工具、工作流程與AI輔助標註的專家,涵蓋LiDAR、點雲與感測器融合。

windags-skills 是什麼?

此技能專注於3D電腦視覺領域,提供LiDAR與點雲標註工具、AI輔助標註工作流程以及感測器融合的專業指導。它詳細比較了2026年的商業與開源標註工具(如BasicAI、Supervisely、CVAT等),並深入解析SAM4D與Point-SAM等最新3D分割模型的架構與應用。此外,技能說明了人機協作標註管線(Human-in-the-Loop)的設計、效率提升策略與品質保證方法,並強調在自駕車、機器人及基礎建設檢測等垂直領域中,使用專門模型(如YOLO、PointPillars)相較於通用視覺語言模型(VLM)的優勢與混合架構的最佳實踐。它支援Claude Code、Cursor、Codex等平台,並透過決策樹與工具選擇矩陣協助使用者快速選定合適方案。

相容平台~Claude Code~Codex CLI~Cursor✓Gemini CLI
npx skills add https://github.com/curiositech/windags-skills/tree/HEAD/skills/3d-cv-labeling-2026

在你喜歡的 AI 中提問

開啟一個已預先載入此 Agent Skill 的新對話。

說明文件

3D 電腦視覺標註專家 (2026)

針對自動駕駛、機器人、基礎設施檢測與地理空間應用中的 LiDAR/點雲電腦視覺,提供 3D 標註工具、AI 輔助標註工作流程與訓練架構的專家指導。

何時使用此技能

✅ 適用於:

  • 選擇 3D 點雲標註工具(BasicAI、Supervisely、Segments.ai、Deepen AI)
  • 實作 SAM4D/Point-SAM 自動標註工作流程
  • 設計人機協作標註管線
  • 感測器融合標註(相機 + LiDAR + 雷達)
  • 訓練架構決策:專業模型 vs VLM
  • 垂直領域 3D 檢測(自動駕駛、檢測、農業、野火)

❌ 不適用於:

  • 不含 3D 背景的 2D 影像標註(請使用 clip-aware-embeddings 或 Label Studio 文件)
  • 一般 ML 模型訓練(請使用 ml-engineer)
  • 不含點雲的影片標註(請使用 computer-vision-pipeline)
  • VLM 提示工程(請使用 prompt-engineer)
  • 攝影測量/3D 重建(請使用地理處理工具)

2026 年工具生態總覽

商業領導者

工具優勢最適用於關鍵 AI 功能
BasicAI一鍵檢測自動駕駛為 AV 微調的預標註模型
Supervisely客製化研發團隊AI 追蹤,2D→3D 單鍵操作
Segments.ai2D+3D 同步機器人感知序列傳播
Deepen AI感測器校準內部感知像素完美多感測器
Dataloop企業級 MLOps大型標註團隊模型輔助 + 點雲專注
Encord完整工作流程多模態專案追蹤 ID 管理
Ango Hub (iMerit)密集標註複雜多模態幀間傳播

開源選項

工具成熟度限制
CVAT穩定僅 3D 邊界框,插值有限
3D BAT良好全環繞標註,半自動追蹤
Label Studio部分 3D更適合多格式,非專用 3D

SAM 在 3D 領域的演進 (2024-2026)

SAM4D (ICCV 2025) - 多模態 + 時序

關鍵創新:統一多模態位置編碼 (UMPE) 將相機與 LiDAR 對齊至共享的 3D 空間。

相機串流 → 特徵提取 → ┐
                                      ├→ UMPE 對齊 → 可提示 3D 分割
LiDAR 串流 → 點編碼     → ┘

資料引擎突破:使用以下方法,自動偽標籤生成速度比人工標註快 100 倍以上:

  1. VFM 驅動的影片遮罩片段
  2. 時空 4D 重建
  3. 跨模態遮罩片段融合

資料集:Waymo-4DSeg(超過 30 萬個相機-LiDAR 對齊的遮罩片段)

Point-SAM (ICLR 2025) - 原生 3D 提示

架構:專為點雲設計的高效變壓器(非從 2D 改編)。

知識蒸餾:2D SAM → 3D Point-SAM,透過資料引擎生成:

  • 部件級偽標籤
  • 物體級偽標籤

基準測試:在室內(ScanNet)與戶外(nuScenes、Waymo)資料集上超越當前最佳結果。

SAMNet++ (2025) - 混合管線

兩階段方法:

  1. SAM 執行非監督式分割
  2. 改編的 PointNet++ 細化語義精度

最適用於:無人機/飛行器工作流程,其中可取得來自 L1 LiDAR + RGB 相機的彩色點雲。


人機協作架構

模型參與標註典範 (2023-2026)

舊方法:人工標註 → 訓練模型 → 部署 新方法:模型輔助 → 人工驗證 → 快速迭代

┌─────────────────────────────────────────────────────────┐
│                    標註管線                              │
├─────────────────────────────────────────────────────────┤
│  原始資料 → AI 預標註 → 人工審查 → 品質檢查             │
│     │           │              │             │          │
│     │     SAM4D/VLM      僅修正         共識抽樣       │
│     │     生成提案       AI 不確定處                    │
└─────────────────────────────────────────────────────────┘

效率提升

方法處理 10k 幀所需時間標註品質
僅人工400 小時95%(專家)
AI 預標註 + 審查50 小時97%(AI+人)
SAM4D 資料引擎4 小時92%(偽標籤)

80/20 法則:約 80% 的 ML 專案時間花在資料準備上。模型參與標註大幅縮短此時間。

品質保證策略

  1. 共識抽樣:對子集進行多位標註員標註,衡量一致性
  2. 主動學習:將不確定的預測路由給專家
  3. 分層審查:第 1 層(關鍵物體)由領域專家驗證,第 2/3 層使用 AI 信心閾值

為何 3D 領域更適合專業訓練而非 VLM

核心取捨

面向專業模型 (YOLO、PointPillars)VLM (GPT-4V、Gemini)
延遲10-50ms(即時)500-2000ms
3D 精度強大的幾何先驗噪雜的文字-3D 對齊
新物體封閉集(訓練過的類別)開放詞彙
運算需求可邊緣部署需 GPU 叢集
幻覺無(確定性)有(安全關鍵風險)
領域偏移表現不佳(霧、夜晚)泛化性較佳

何時使用哪一種

使用專業模型時機:

  • 需要即時推論(自動駕駛車輛、機器人)
  • 已知物體類別(基礎設施缺陷、作物類型)
  • 安全關鍵部署(無法容忍幻覺)
  • 邊緣部署(無人機、嵌入式系統)

使用 VLM/基礎模型時機:

  • 探索新領域的零樣本學習
  • 生成訓練資料(弱標籤)
  • 開放詞彙需求(「找任何受損物品」)
  • 領域適應引導

混合架構 (2025+ 最佳實踐)

                    ┌───────────────────────┐
                    │    VLM(慢思考)       │
                    │  • 場景理解            │
                    │  • 開放詞彙            │
                    │  • 異常檢測            │
                    └──────────┬────────────┘
                               │ 高層級背景資訊
                               ▼
┌──────────────────────────────────────────────────────────┐
│              專業檢測器(快思考)                          │
│  • 即時推論 (YOLO、PointPillars、CenterPoint)            │
│  • 已知物體檢測與追蹤                                    │
│  • 安全關鍵決策                                         │
└──────────────────────────────────────────────────────────┘

範例:

  • VOLTRON:YOLOv8 + LLaMA2 用於危害識別
  • DrivePI:點雲 + 多視角 + 語言指令 (0.5B Qwen2.5)

垂直領域特定訓練架構

基礎設施檢測

物體:電線桿、絕緣子、導線、植被、損壞類型 感測器融合:RGB + 熱成像 + LiDAR 訓練資料需求:

  • 熱異常樣本(不同溫度)
  • 損壞分類(裂縫、腐蝕、鏽蝕等級)
  • 植被淨空測量

架構:

LiDAR → 點雲編碼器 → ┐
熱成像 → 2D 編碼器  → ├→ 融合 → 多任務頭
RGB → 2D 編碼器    → ┘          ├→ 物體檢測
                                         ├→ 缺陷分類
                                         └→ 淨空迴歸

自動駕駛

物體:車輛、行人、自行車騎士、交通號誌、車道標線 關鍵需求:時序一致性(跨幀追蹤 ID) 訓練資料需求:

  • 長尾場景(緊急車輛、動物、碎片)
  • 惡劣天氣(霧、雨、雪、夜晚)
  • 邊緣案例(施工區域、事故)

架構:使用 BEV(鳥瞰圖)表示的 CenterPoint、PointPillars 或體素檢測器。

農業/野火

物體:作物行、樹冠高度、燃料負載、火勢蔓延邊界 感測器融合:RGB + 多光譜 + LiDAR 訓練資料需求:

  • 作物生長階段
  • 疾病/蟲害視覺特徵
  • 來自 LiDAR CHM(樹冠高度模型)的燃料負載密度

為何不只用 VLM? VLM 無法:

  • 測量精確高度(LiDAR 迴歸)
  • 在超光譜波長下分類
  • 維持處方圖所需的空間精度

常見反模式

反模式:「什麼都用 SAM」

新手想法:「SAM 能分割一切,所以我直接在 LiDAR 資料上執行它」

現實:

  • SAM 1/2 是 2D 模型——它們不懂 3D 幾何
  • 點雲需要 Point-SAM 或 SAM4D
  • 直接套用會產生不含幾何先驗的噪雜遮罩

正確方法:使用 Point-SAM 進行原生 3D 處理,或投影至 2D 讓 SAM 處理後再回升至 3D。

反模式:跳過人工驗證

新手想法:「AI 預標註有 95% 準確率,我們可以跳過審查」

現實:

  • 10 萬個物體中 5% 的錯誤 = 5,000 個錯誤標籤
  • 錯誤在邊緣案例中會疊加(正是你需要的準確之處)
  • 模型會學習複製標註錯誤

正確方法:第 1 層(安全關鍵)始終由人工驗證。第 2/3 層使用信心閾值。

反模式:VLM 用於即時推論

新手想法:「GPT-4V 可以辨識我照片中的損壞」

現實:

  • 每幀 500-2000ms 延遲
  • 無法在邊緣裝置上執行
  • 在安全關鍵背景下有幻覺風險

正確方法:VLM 用於資料生成/探索,專業模型用於部署。

反模式:單一模態訓練

新手想法:「LiDAR 對 3D 檢測來說就夠了」

現實:

  • LiDAR:精確幾何,無色彩/紋理
  • 相機:豐富語義,無深度
  • 融合在 mAP 上比單一模態好 5-15%

正確方法:從第一天起就使用感測器融合。SAM4D 顯示融合偽標籤 > 單一模態。


決策樹:選擇你的方法

                        需要即時推論嗎?
                              /                  \
                           是                    否
                            |                      |
                   使用專業檢測器           這是探索性任務嗎?
                   (YOLO, CenterPoint)        /        \
                            |                 是         否
                   有標註資料嗎?         使用 VLM     使用 SAM4D
                      /        \         進行零樣本   生成偽標籤
                   是          否                        |
                    |            |
             訓練模型    使用 SAM4D/
                         Point-SAM 自動標註

工具選擇決策矩陣

需求建議工具
大規模自動駕駛Deepen AI 或 BasicAI
研發/研究靈活性Supervisely 或 Segments.ai
多模態(相機+LiDAR+雷達)Ango Hub 或 Dataloop
自託管/開源CVAT + 3D 外掛或 3D BAT
機器人感知Segments.ai (2D+3D 同步)
預算有限Label Studio + 自訂腳本

參考資料

  • /references/sam4d-architecture.md - SAM4D UMPE 與資料引擎深入探討
  • /references/tool-comparison-matrix.md - 所有工具的詳細功能比較
  • /references/hybrid-architecture-examples.md - VOLTRON、DrivePI 實作模式
  • /references/vertical-training-recipes.md - 基礎設施、自動駕駛、農業細節

來源

Individual skills in this repo

This repo contains 20 individual skills — each has its own dedicated page.

curiositech/windags-skills

Expert in 2000s-era music visualization (Milkdrop, AVS, Geiss) and modern WebGL implementations. Specializes in Butterchurn integration, Web Audio API AnalyserNode FFT data, GLSL shaders for audio-reactive visuals, and psychedelic generative art. Activate on "Milkdrop", "music visualization", "WebGL visualizer", "Butterchurn", "audio reactive", "FFT visualization", "spectrum analyzer". NOT for simple bar charts/waveforms (use basic canvas), video editing, or non-audio visuals.

curiositech/windags-skills

Expert legal research agent for finding and scraping expungement data state by state. Knows authoritative sources, URL patterns, Firecrawl configuration, and 2026 legal landscape.

curiositech/windags-skills

Implement WCAG 2.2 AA/AAA compliance with automated testing, keyboard navigation, screen reader support, and focus management. Activate on: accessibility audit, WCAG compliance, keyboard navigation, screen reader, aria attributes, axe-core, focus trap. NOT for: design-level accessibility review (use design-accessibility-auditor), color contrast only (use css-in-js-architect).

curiositech/windags-skills

Time-blind friendly planning, executive function support, and daily structure for ADHD brains. Specializes in realistic time estimation, dopamine-aware task design, and building systems that actually work for neurodivergent minds.

curiositech/windags-skills

Designs digital experiences for ADHD brains using neuroscience research and UX principles. Expert in reducing cognitive load, time blindness solutions, dopamine-driven engagement, and compassionate design patterns. Activate on 'ADHD design', 'cognitive load', 'accessibility', 'neurodivergent UX', 'time blindness', 'dopamine-driven', 'executive function'. NOT for general accessibility (WCAG only), neurotypical UX design, or simple UI styling without ADHD context.

curiositech/windags-skills

>- Apply crisis decision-making research to agent routing, uncertainty triage, and coordination failure analysis in time-pressured systems. Use when diagnosing handoff failures, analytical paralysis, or expert judgment under incomplete information. NOT for routine coding, simple CRUD design, or static single-agent tasks with complete information.

curiositech/windags-skills

Extend and modify the admin dashboard, developer portal, and operations console. Use when adding new admin tabs, metrics, monitoring features, or internal tools. Activates for dashboard development, analytics, user management, and internal tooling.

curiositech/windags-skills

Conversation patterns and interaction protocols for multi-agent systems. Covers request/response, pub/sub, blackboard, delegation chains, debate, critique, consensus, fan-out/fan-in, supervisor-worker, and peer negotiation. Deep analysis of AutoGen conversation patterns, CrewAI delegation, LangGraph state passing, and FIPA-ACL performatives. Teaches how to design what agents say to each other and in what order. Activate on: "agent conversation", "agent protocol", "multi-agent debate", "agent delegation", "supervisor worker pattern", "agent voting", "consensus protocol", "fan-out fan-in", "agent negotiation", "blackboard pattern", "agent dialogue", "conversation topology", "agent handoff". NOT for: wire format or serialization (use agent-interchange-formats), orchestration infrastructure (use agentic-infrastructure-2026), single agent behavior (use agentic-patterns).

curiositech/windags-skills

Meta-agent for creating new custom agents, skills, and MCP integrations. Expert in agent design, MCP development, skill architecture, and rapid prototyping. Activate on 'create agent', 'new skill', 'MCP server', 'custom tool', 'agent design'. NOT for using existing agents (invoke them directly), general coding (use language-specific skills), or infrastructure setup (use deployment-engineer).

curiositech/windags-skills

AI-powered calendar management and agent-based scheduling coordination. Covers calendar APIs (Google Calendar, CalDAV/iCal), AI scheduling assistants (Reclaim, Clockwise, Motion, Cal.com), building custom calendar agents with MCP, multi-calendar merging, timezone management, focus block protection, meeting fatigue detection, and agent-to-agent meeting negotiation protocols. Activate on: "calendar agent", "AI scheduling", "calendar coordination", "meeting scheduling", "calendar API", "focus time protection", "calendar optimization", "Google Calendar MCP", "Reclaim", "Clockwise", "Motion", "Cal.com", "smart scheduling", "calendar-aware agent", "timezone scheduling", "agent negotiation meetings". NOT for: manual calendar UI component design (use form-validation-architect), project management scheduling or Gantt charts (use project-management-guru-adhd), general time-tracking or pomodoro apps (use adhd-daily-planner for time-awareness), building the agent itself from scratch (use agent-creator).

curiositech/windags-skills

Build and adopt production AI agent infrastructure in 2026. Covers framework selection (LangGraph, CrewAI, AutoGen, MCP), orchestration patterns, evaluation, observability, memory systems, and tool use. Also covers the SOCIAL dimension: how to sell agent infrastructure internally, change management, measuring ROI, building trust in autonomous systems, and scaling adoption across teams. Activate on: "agent infrastructure", "agent framework comparison", "which agent framework", "sell AI tools internally", "agent adoption", "agent observability", "agent evaluation", "MCP architecture", "agentic mesh", "enterprise AI agents", "AI change management", "agent ROI". NOT for: building specific agents (use ai-engineer), designing agent behavior patterns (use agentic-patterns), prompt tuning (use prompt-engineer).

curiositech/windags-skills

Fundamental patterns for effective agentic behavior. Teaches decomposition, tool orchestration, error recovery, context management, quality self-assessment, and knowing when to stop. Model-agnostic principles that make any agent more effective regardless of domain. Activate on: "how should I structure this agent", "agentic workflow", "agent patterns", "multi-step task", "tool orchestration", "/agentic-patterns", "decompose this", "agent best practices", "chain of actions", "when should the agent stop", "agent loop design". NOT for: creating agent infrastructure (use agent-creator), building DAGs (use windags-architect), specific tool implementation.

curiositech/windags-skills

Automated discovery and matching of agent skills for dynamic task routing and capability assessment

curiositech/windags-skills

Cryptographic security for agentic systems — zero-trust agent networking, signed message envelopes (JWS/JWE), capability-based security (ocaps), Merkle tree audit trails, WASM sandboxing, and formal verification. Covers CLI dev tool security, mTLS between agents, permission boundaries (least privilege for AI agents), and supply chain security for skills/plugins. Activate on: "agent security", "zero trust agents", "secure agent communication", "capability-based security", "ocap", "signed messages between agents", "agent audit trail", "sandbox agent execution", "agent permissions", "mTLS agents", "cryptographic verification", "agent supply chain", "OWASP agentic", "prove agent did X", "tamper-proof agent logs". NOT for: application-level SAST scanning (use security-auditor), network firewall rules (use infrastructure), SOC2/HIPAA compliance (organizational), or prompt injection defense (use prompt-engineer).

curiositech/windags-skills

Data structures and serialization formats for agent-to-agent communication. Covers message envelopes, structured output schemas, capability declarations, task handoff payloads, error/retry signaling, and context windows as data structures. Deep comparison of A2A protocol, MCP, OpenAI function calling, and LangChain message types. Teaches when to use rigid schemas vs free-form with validation, typed vs untyped, streaming vs batch. Activate on: "agent message format", "agent communication schema", "agent-to-agent protocol", "A2A protocol", "MCP message format", "structured output for agents", "agent interop", "interchange format", "agent serialization", "task handoff format", "capability declaration". NOT for: what agents say to each other (use agent-conversation-protocols), orchestration topology (use multi-agent-coordination), building agent infrastructure (use agentic-infrastructure-2026).

curiositech/windags-skills

Logic-based agent programming language implementing BDI architecture for practical autonomous agent development

curiositech/windags-skills

>- Design AgentSpeak(L)-style BDI agents with context-guarded plans, selection functions, and intention stacks. Use for interruptible autonomy, agent policy, and multi-agent orchestration in dynamic environments. NOT for simple rule engines, static planners, or centralized workflows.

curiositech/windags-skills

Foundational concurrent computation model where actors communicate exclusively through asynchronous message passing

curiositech/windags-skills

Build production-ready LLM applications, advanced RAG systems, and intelligent agents. Implements vector search, multimodal AI, agent orchestration, and enterprise AI integrations. Use PROACTIVELY for LLM features, chatbots, AI agents, or AI-powered applications.

curiositech/windags-skills

license: Apache-2.0 NOT for unrelated tasks outside this domain.

相關技能