AbstractsComputer Science

This thesis is dedicated to the implementation of a machine vision algo-rithm on a Raspberry Pi Microcomputer. The commissioning client Teo Kanniainen conducted a research project on the subject of arranging low-cost apple moth inspection on his garden. The research resulted in production of a device collecting pictures of insects in order to process these images later. The contact person Markku Kippola discussed the issues of improved functionality and cost, so the overall aim of this thesis thus was to develop a reliable cheap alternative to Kanniainen’s device. The end product of the this project is an autonomous embedded system for inspection and reporting of apple moth, which functionalities can be further extended to perform other tasks, such as dispatching pesticide control. The designed system harnesses the powerful modularity of the Python scripting language and the OpenCV machine vision framework and utilises the widely used modular microcomputer, the ‘Raspberry Pi’. The reader of this thesis will gain understanding of how to program a Raspberry Pi microcomputer and will become familiar with algorithms used in machine vision that take advantage of both vision and machine learning capabilities. The thesis also provides a real life python code and demonstrates how to install and program machine vision applications with the OpenCV library. This document describes the issues faced and the solutions found in this particular case.