Printed circuit boards are designed in high density, small size and compact layout. It is obvious that faults like open hole, breakout and missing component may occur at the time of manufacturing. The electrical short circuit, dirt, corrosion also leads to problems like crack , burnt components in a PCB during usage. Identifying the faults in a non-contact method through image processing makes the work easy and safe. This paper proposes an algorithm of comparing reference PCB image with faulty boards image. Morphological operations Opening ,Closing, Bottom-hat filtering are applied on the images of test boards and finally subtracted from the reference image to locate the faults. This method identifies cracks, open hoe, missing components and short circuit faults in less time. © 2019 Mattingley Publishing. All rights reserved.