Abstract ▼ AI-Helper

The goal of robust methods in computer vision is to extract all the information necessary to solve a given task while discarding everything that is not needed. The tasks can be very simple or very complex, but in real-life applications, a robust procedure is always required. In the end, the performa...

The goal of robust methods in computer vision is to extract all the information necessary to solve a given task while discarding everything that is not needed. The tasks can be very simple or very complex, but in real-life applications, a robust procedure is always required. In the end, the performance of a machine for solving a vision problem will be judged against that of human observers performing the equivalent task. Since we know that the human visual system works in a much more sophisticated manner than the present day computer vision systems, this ultimate goal is still far away. Nonetheless, the aim of robust computer vision systems has always been to emulate human vision-like behavior in the presence of noise. Furthermore, the robust algorithms should be independent of user inputs up to quite a large extent. However, contrary to this, almost all state-of-the-art robust estimation algorithms are dependent on the user for providing some information about the underlying characteristics of the data on which the algorithm operates. Often times, it is hard for the user to supply such information to the algorithm. The work presented here focuses on developing robust algorithms for computer vision that can estimate the underlying model in the data without any sort of user intervention. We present some interesting applications in inspection of mechanical parts. In first part, we present a novel radius measuring algorithm which has a capability to effectively measure the radius of small circles under a lightening noise as well as jig/fixture positioning error. Moreover, our algorithm can be applied to calculate the center point of any polygonal shapes that is very difficult to be implemented by regular circle detection algorithm. In the algorithm we adopt an iterative fuzzy searching method that searches the center of any shapes by fuzzy parameters of distance and orientation from initial search point, which can reduce computational complexity significantly and accurately detect the center point under severe noisy environment. The proposed algorithm has been implemented and tested on both synthetic and real-world part images, and the performance is compared to popular circle detection algorithms to prove the accuracy and effectiveness of the algorithm. From the comparison it is concluded that the proposed algorithm gives an excellent performance in measuring accuracy and computation time. In the second part, we present a new algorithm for inspection of grid marking application. Recently, many researchers interest in GPU powered applications and algorithms to use its ultimate parallel computation power. Using contour information we introduce a new growing snake algorithm for parallel ellipse detection for formability evaluation test. Perspective view of the deformed ellipse is hard (almost impossible) to extract real ellipse parameter using 2D image. Then we enhance our ellipse measurement result using stereo vision based depth measurement. The formability of the sheet metal is then evaluated using limit forming diagram. In this paper we utilize AZ31 magnesium alloy sheet metal with circular grid pattern. With this particular sheet metal characteristic we provide experimental result using a forming limit diagram. The forming limit diagram is provided by strain limit test experiment. We inform that our algorithm achieved very high accuracy with high speed evaluation with assist of GPU.

주제어