AbstractThe outlook examination in fabrication of plastic products in the past has been performed mainly by visual inspection. The automation of the procedure is now considered to be important. This paper discusses a simple automatic detection method for defects on the surface of plastic electronic components such as integrated circuits. As the first step, optical properties of the defect such as flaw, pin‐hole and bulge on the surface of the plastic components are discussed. This indicates problems in the detection of those defects. Then the illumination method using a planar light source is discussed to obtain a stable image against surface inclination. The procedure for practical detection of defects is developed in this paper as follows. First, to separate the image of the component surface from the background, the projection distribution of the gray‐level histograms is constructed. The boundary of the component surface is extracted based on the histogram. Then the surface image is partitioned into subareas. The gray‐level histogram is constructed for each partitioned subarea. Based on the maximum and the most‐frequent value of the gray‐level histogram, the existence or nonexistence of a defect in the subarea is determined. The experiment for the defect detection was performed using typical plastic components which are determined good or no‐good. It is demonstrated that a correct defect detection can be performed for all kinds of components.