An implementation for round object detector in the image by using the inner product detection methodology is proposed. Applying the geometric characters of the circumferential angle, the inner product detection is developed to determine whether there are round objects in the image. If the round objects exist, the center coordinates and the radii of the corresponding circles are evaluated, and give a description of the objects. To overcome the digital error problems in image processing, we present the threshold value adjustment method for the circumferential angle which the threshold value can be adjusted automatically based on the size of the radius. Computer simulations for a detection of the images with several round objects were performed to examine the effectiveness of the proposed inner product detection methodology. From the results of simulation, the scheme illustrates the good performance of the round objects detection.