Defect Shape Matching Simulation Results

74

3.3 Simulation Results

Each of simulation development is tested using sample object to configure whether the system capable in achieving the objectives of the research. Flexibility of the system is stressed in this research in order to produce a system that able to complete its task with high efficiency. The testing is conducted using webcam 12.1 megapixels instead of CCD camera. The purpose is to clarify the algorithm efficiency and accuracy using low quality image before testing using real application.

3.3.1 Defect Shape Matching

In this system, object pyramid is created using plain cardboard with base dimension 15cm x 15cm and height nearly about 11.82cm from base. 15 tested pyramids used in testing the availability of the algorithm in classifying pyramid object. Three factors that play a major role in determining the success of the best available is the object scale, rotation of objects and lights from environment. These three factors are tested for their efficacy to carry out three experiments respectively using the three most important factors in determining the level of matching object model used. The first experiments carried out at a height of 60cm from the webcam to the base where it is regarded as a position scale. The second and third experiments were conducted at the height of 40cm and 80cm respectively. Experiment is conducted in the dim light lab to prevent the emission of light is so bright that could affect the results. The processed images will undergo 6 matching process from 0 degrees to 360 degrees 0 360 degrees, 60 degrees, 120 degrees, 180 degrees, 240 degrees and 300 degrees. All images are gone through the matching process to determine if the system is able to get good results even if the image is rotated from the original 75 position. In addition, each image will also be tested with the light-controlled webcam with no lighting, half lighting and full lighting. All the data are recorded in Table 3.4, Table 3.5 and Table 3.6, as shown. Table 3.4: Experiment 1; Webcam Position = 60 cm from base Figure No Lighting Webcam Half Lighting Webcam Full Lighting Webcam Pointing Result 1  0 o 360 o Score=0.976433 Score=0.980151 Score=0.982591 Success 2  60 o Score=0.911244 Score=0.914525 Score=0.891157 Success 3  120 o Score=0.865419 Score=0.868720 Score=0.892017 Success 4  180 o Score=0.919319 Score=0.928821 Score=0.957525 Success 5 240 o Score=0.914869 Score=0.891402 Score=0.911028 Success 6  300 o Score=0.878182 Score=0.869127 Score=0.902388 Success Total Score=91.09 Scale=1.002931 Score=90.88 Scale=1.004977 Score=92.28 Scale=1.007235 76 Table 3.5: Experiment 2; Webcam Position = 40 cm from base Figure No Lighting Webcam Half Lighting Webcam Full Lighting Webcam Pointing Result 1  0 o 360 o Score=0.897040 Score=0.883717 Score=0.873799 Success 2  60 o Score=0.828437 Score=0.87104 Score=0.874709 Success 3  120 o Score=0.806566 Score=0.845248 Score=0.836497 Success 4  180 o Score=0.921424 Score=0.937709 Score=0.873799 Success 5 240 o Score=0.886828 Score=0.856704 Score=0.832641 Success 6  300 o Score=0.825031 Score=0.864090 Score=0.848658 Success Total Score=86.09 Scale=1.475830 Score=87.64 Scale=1.479893 Score=85.67 Scale=1.484594 Table 3.6: Experiment 3; Webcam Position = 80 cm from base Figure No Lighting Webcam Half Lighting Webcam Full Lighting Webcam Pointing Result 1  0 o 360 o Score=0.810524 Score=0.810285 Score=0.873630 Success 2  60 o Score=0.802063 Score=0.771798 Score=0.848408 Success 3  120 o Score=0.851231 Score=0.793409 Score=0.864623 Success 4  180 o Score=0.852781 Score=0.835162 Score=0.858725 Success 5 240 o Score=0.800803 Score=0.808714 Score=0.852678 Success 6  300 o Score=0.790774 Score=0.806396 Score=0.900508 Success Total Score=81.80 Scale=0.754380 Score=80.43 Scale=0.762243 Score=86.64 Scale=0.758876 77

3.3.2 Defect Shape Pointing