Uncertainty Analysis Theory

1255 Words6 Pages
Abstract — This article presents the stages of planning and analysis used to determine the influence of certain factors in controlling geometric figures dimension by computer vision. It was used as basis techniques used as image processing to inspect objects, in order to reduce measurement errors in existing manual process. It uses the 3k factorial design where it is assessed that it´s possible of processing a digital image of a particular geometric figure whose dimensions are known. It will determine the measurement uncertainty by Monte Carlo (MMC). Finally, it was possible to reject the null hypothesis that the size does not influence the result at dimensional figures. Keywords: ANOVA, Dimensional Control, Experimental Procedure, Monte…show more content…
B. Methods for Uncertainty Analysis The uncertainty analysis can be performed by different methods. One method is the Guide of Uncertainty Measurement (GUM) [5]. This method consists in showing that the uncertainty of a measurement can incorporate various sources of uncertainties. These uncertainties are from random errors and systematic errors [3]. GUM guide usage begins in the measurement equation, which should include all relevant contributions for calibration or testing. It is estimated the overall uncertainty by law of propagation of uncertainty through the identification and quantification of the uncertainty of individual influence factors [5, 6]. During the evaluation of measurement uncertainty, it may be that source of error go unnoticed, due to the limited knowledge of the evaluator. The Monte Carlo method uses the concept of spreading to entire probability distribution of the input. There is supplement to the Guide to Measurement Uncertainty (GUM), which deals with the simulation of the Monte Carlo for calculating measurement…show more content…
In this way, the environment where the design has been developed is according to the standard. The camera used was Cyber-shot DSC-W35 Sony. The chosen resolution was seven Mega Pixel with the goal of having the highest number of digital image information. For the measurements, it was used a ruler graduated in millimeters. For digital image processing it was used a laptop processor Intel Pentium M, 1.73 GHz, 503 MB de RAM. The digital processing of the images used the software compiler Borland C++ Builder 5, together with the libraries of LiLi. This library was developed with the goal of providing a unique and efficient handling and processing images. Aiming portability, the library was written in the language ANSI/ISO C [2]. The results of analysis can be performed using hypothesis testing [1]. This test assesses the influence of the factors and their interactions on the response variable. The test used to evaluate the results is the F-type test statistic, which tests equality between averages. The assumptions used in the work
Open Document