Reducing the Probability of Large Errors in Position Detection

D. Karimzadgan and K. Khalili (Iran)


: Object Detection, Accuracy Improvement, Machine Vision, Flexible Assembly, Printed Circuit Board Assembly.


: Locating objects within a tolerance are one of main issues in applications such as assembly. One way of reducing the probability of large errors occurring when locating objects is to repeat measurements and take the average of repeated readings as the true value of the measurand. The current work is based on the idea of using multi-point measurement instead of multi measurement of a point. In many applications where the absolute position of a point is to be determined, there are known relationships between the point of interest and other points. This information can be used to locate the point of interest with higher accuracy. This paper describes a simple method which makes use of some prior knowledge about the object being viewed resulting in reduction in probability of large errors occurring when locating objects. Previous works developed a Geometric Method (GM) and a Genetic Algorithm Method (GAM). Both GM and GAM have their limitations. The method presented in this paper can overcome some of these problems.

Important Links:

Go Back