Machine vision technology has permeated many areas of industry, and automated inspection systems are playing increasingly important roles in many production processes. Electronic manufacturing is a good example of the integration of vision based feedback in manufacturing and the assembly of surface mount PCBs is typical of the technology involved. There are opportunities to use machine vision during different stages of the surface mount process. The problem in the inspection of solder joints on surface mount printed circuit board is much more difficult than many other inspection problems. In this thesis, an approach for inspecting surface mounted integrated circuits (SMICs) is presented. It is based on the variance of intensity values of pixels in an image. This method is able to cope with 4 kinds of soldering defects in SMICs. A set of modules for the system is proposed. The computer program which performs the image processing and analyzing has been written in C. It has been linked with a number of image processing routines from MAVIS1 to perform some image processing tasks, and the result is a compact executable module which works under MS-DOS2 3.30.