CrackDect: Detecting crack densities in images of fiber-reinforced polymers

Abstract

CrackDect is a tool to detect cracks in a given direction from a series of images. It is specialized to detect multiple matrix cracks in composite laminates to yield the crack density but can also be used as a general line detection. The package is written in Python, and includes classes and functions to efficiently handle large image stacks, pre-process images and perform the crack detection. Due to its modular structure it is easily expandable to other crack detection or feature recognition algorithms. Pre-processing of whole image stacks can be customized to account for different image capturing techniques. Since image processing tends to be computational and memory expensive, special focus is put on efficiency.

Keywords

• Crack detection
• Image processing
• Fiber-reinforced polymers
• Crack density
• Python