course-image-parser

Extract course geometry and design parameters from robotrace diagram images into JSON.

7|2|Updated May 31, 2026

Search Tags:#automation #geometry #computer vision #image analysis #cad #robotrace

Ready to run

npx skills add https://github.com/kyamashita5/robotrace_course_cad --skill course-image-parser

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: course-image-parser
Source: https://github.com/kyamashita5/robotrace_course_cad/tree/main/.agents/skills/course-image-parser
Command: npx skills add https://github.com/kyamashita5/robotrace_course_cad --skill course-image-parser

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires uv, and includes scripts (resource) components.

What problem does it solve?

This skill automates the labor-intensive process of converting hand-drawn or printed robotrace course diagrams into precise, CAD-ready JSON files, eliminating manual geometry estimation and coordinate entry.

Core Features & Use Cases

Automated Geometry Extraction: Detects board contours, start/goal positions, and centerline paths from course images.
Intelligent Design Consolidation: Integrates AI-assisted OCR of red design annotations with scripted support-circle and slalom template detections.
Use Case: When designing a new course for a competition, upload the course diagram image and use this skill to generate a validated helper-circle JSON model that is ready for final review and refinement in the Robotrace Course CAD environment.

Quick Start

Use the course-image-parser skill to analyze the course diagram in data/my-course.png and generate the initial CAD JSON candidates.

Frequently Asked Questions about course-image-parser

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I extract course geometry from a robotrace diagram image?▼

To extract course geometry from a robotrace diagram image, the skill automates board normalization, centerline tracing, and helper-circle candidate generation using computer vision pipelines. This converts visual course data directly into structured JSON design artifacts for CAD modeling.

What is the best way to convert printed robotrace course images into CAD-ready files?▼

The best way to convert printed robotrace course images into CAD-ready files is through automated image analysis that detects board contours, start/goal positions, and centerline paths. It eliminates manual geometry estimation by integrating OCR of design annotations with scripted support-circle detection.

Can I automate the extraction of start and goal positions from a course diagram?▼

Yes, you can automate the extraction of start and goal positions from a course diagram. The workflow applies image processing and computer vision techniques to detect board contours and centerline paths, integrating AI-assisted OCR for red design annotations to consolidate the extracted data.

Do I need uv to run the course image parser?▼

Yes, you need uv to run the course image parser. It is a required dependency for executing the scripts that handle the end-to-end workflow of board normalization, centerline tracing, and helper-circle candidate generation from your robotrace course diagram images.

Does the course geometry extraction handle hand-drawn design annotations?▼

Yes, the course geometry extraction handles hand-drawn design annotations through AI-assisted OCR. It specifically reads red design annotations on the diagram and integrates them with scripted support-circle and slalom template detections to generate a validated JSON model.

What are the limitations of automated centerline tracing from course images?▼

The limitations of automated centerline tracing from course images include the need for final review and refinement in the Robotrace Course CAD environment. The generated helper-circle JSON candidates serve as an initial model, requiring validation against the original visual course data.