Marker-Based Tracking Under Extreme Illumination : Feasibility Study And Pilot Testing

Fauzi, Muhamad Afiq Afifi (2018) Marker-Based Tracking Under Extreme Illumination : Feasibility Study And Pilot Testing. Project Report. UTeM, Melaka, Malaysia. (Submitted)

Text (24 Pages) Marker-Based Tracking Under Extreme Illumination ; Feasibility Study And Pilot Testing.pdf - Submitted Version Download (399kB)

Abstract

The tracking system becomes a necessary need in the modern illumination system in the surgical room. This modern illumination system can replace the shortcomings of manually or old illumination system. This shortcomings including having to adjust surgical lamp manually and the presence of shadow due to obstacle that influence the amount of light reaching to the surgical area. To make this tracking system become efficient, a correct characterization of marker need to be identified. For this project, the characteristic of marker were proposed based on color and shape. This project was proposed to overcome two major problem which were the presence of shadow that inhibits transmission light and the changes of properties on the objects due to high illumination. So this project has focused on a few goal. Firstly was to analyze suitable object for object recognition. Both color and shape characteristic were used to recognize the marker. Black balls of different sizes were chosen due to properties of shape and color which were black and circular. Next was to locate the marker’s position and to evaluate the precision of marker’s position localization. Three different experiments were conducted by varying the brightness of light, size of object and the position in camera space in order to evaluate the tracking capability. Raspberry Pi with Pi camera were chosen to realize these experiments. Algorithms were written in Python with help from OpenCV libraries. While surgical lights can reach 45000 lux, these experiments were conducted at 5000 lux at maximum due to limited illumination capacity. Therefore the results may only be applicable to lower specification of surgical light such as illumination for dental surgery. It was found that marker with a diameter of 9.3 cm was consistently tracked within all quadrants in the experiment and at distance 100 cm, the error produce in marker detection was the least. The bigger the size was the better for the object to be detected in camera space and object was less susceptible to error when viewed from a longer distance since the apparent circle was smaller and the discrepancy of the approximation of the circle outer edge was minimal.

Actions (login required)

View Item