Abstract
Swiftlets are small insectivorous birds which breed throughout Southeast Asia and the South Pacific. Among various species of swiftlets, only a few are notable to produce edible bird’s nest (EBN) from the secreted saliva during breeding seasons. It is also one of the demanding food product in food production industry of South East Asia. Government looking into ways to improve this industry to boost the economy. However, one of the major difficulties faced in processing the bird’s nest is to remove its impurities, also known as dirt. Conventional way of cleaning the EBN is way too inefficient and time-consuming. Moreover, it requires large amount of labour forces in cleaning a small amount of bird’s nest. This paper presents an automated system which combines machine vision and cartesian robotic system in order to achieve better efficiency in cleaning the EBN. Prior to the system software, Microsoft Visual Studio and OpenCV libraries are implemented to process the image. For the hardware part, cartesian robotic system and ultrasonic cleaner is implemented with the aid of high resolution greyscale camera. Image processing utilises the morphology of thresholding and binarization combined with BLOB analysis to locate the locations of all the dirt. As a result, the tiny dirt present in the EBN is inspected and a cleaning mechanism using cartesian robotic system and ultrasonic cleaner are designed to clean the dirt.
Index Terms—Edible Bird’s Nest (EBN); Cartesian Robotic System; Thresholding; Blob Analysis; Binarisation; Machine Vision; Ultrasonic Cleaner.