Circuit World

Purpose – The purpose of this paper is to study the application of advanced computer image processing techniques for flaw detection on flexible printed circuit (FPC) solder.

Design/methodology/approach – Texture directionality feature is obtained based on texture gradient, contour's position is extracted and directionality information obtained through analyzing the distribution of directionality. Contour similarity function is established to filter out false contour and keep proper contour, and the solder's location work is accomplished based on reversed contour. After that, a combination of grey and texture gradient's value deviation from reference value is utilized to reflect and describe texture on the solder's surface. Flaw can be distinguished from homogeneous texture background.

Findings – The method has been applied to the inspecting system and achieved a higher accuracy and a lower false defect rate. It demonstrates that the method can detect flaws efficiently and effectively.

Research limitations/implications – Although the work on FPC solder's location and flaw detection is presented, defective classification is not involved that is also very important content for inspection.

Originality/value – The paper provides a new way to locate solder based on directionality. The method not only extracts contour feature but also gains directional parameters to help realize accurate location, especially for some solders that are deformed to some extent. Entropy statistic based on distribution of grey and texture gradient is proposed to describe and measure solder's surface texture. The new algorithm performs stably and efficiently and is fit for practical application.