Various concentrated works have been done in the area of computational vision regarding the circle and texture detections. Detection of circles in images can be beneficial for PCB components industries for the detection of capacitors in printed circuit boards, also for medicine in the detection of red cells, white blood cells, and leukocytes, and for applications which requires precision and assignments regarding the detection of circles in a digital image. In this work is utilized a benchmarking of images to detection circle boards of different radio values for the comparison with the work [1] of this article. The benchmarking of images is composed of five main images that are tested in the algorithm of detection of circles in MATLAB with different values of radio for each image. The results appoint an enhancement of 300 % concerning the algorithm proposed in work [1] showed in this article. In this work also would be plotted graphs concerning the accuracy of the new proposed algorithm with relation to the algorithm proposed in work [1], indicating better results concerning the GUI interfaces and capacity of detection circles.