Evaluation of machine vision system performance under different illumination geometry

Abstract

Light geometry is a very important role for machine vision but there a limited study in evaluating machine vision performance under different geometry of vision illumination has been reported especially plant sample for agricultural application. With this experiment, a few light techniques or geometry was tested to obtain which geometry suitable in agriculture especially for used in green plants. In this experiment, 3 light geometries selected were the experiment is Dome light, Square light and Line square light. This three geometry was tested under y different height and angle to know which height and angle that suitable for green plant fields. To know which lighting that suitable for green plants, objective of this research must be completed get the best result. After all experiments were done, the image acquired from Dome light a bit filtering because dome light is diffuse cloudy type that make the light no directly interface with sample and. For Square light also diffuse type, but the image look clearly and the RGB value get from image also show the greenness of plants. Lastly, Line square light is a direct light type that makes image capture clearer, but only for angle 75° and 90°, for angle 35° and 45° the image look darker because this angle produced a lot shadow and that make image seeing darker than other people.