Multidirectional Scanning Model, MUSCLE, to Vectorize Raster Images with Straight Lines.

Area of Science:

• Computer Vision
• Image Processing
• Computational Geometry

Background:

• Automatic vectorization of raster images is essential for digital archiving and editing.
• Existing methods may lack accuracy or flexibility for diverse technical drawings.
• Straight line extraction from raster data remains a challenging problem in image analysis.

Purpose of the Study:

• To develop and present a novel algorithm for the automatic vectorization of raster images containing straight lines.
• To enable user-defined criteria for customizable vectorization processes.
• To demonstrate the model's applicability across various technical drawing types.

Main Methods:

• The MUSCLE (Multidirectional Scanning for Line Extraction) model employs line thinning and neighborhood analysis algorithms.
• User-defined parameters allow for tailored vectorization based on specific requirements.
• The algorithm was implemented using computer programming and tested on diverse raster datasets.

Main Results:

• The MUSCLE model successfully vectorized raster data, including township plans, maps, and architectural drawings.
• Performance was validated against established vectorization software (WinTopo, Scan2CAD).
• The model demonstrated both speed and accuracy in the vectorization process.

Conclusions:

• MUSCLE provides an effective and efficient solution for automatic straight line vectorization from raster images.
• The model's flexibility in handling user criteria enhances its utility for various applications.
• Further development and testing on more complex datasets are warranted.