Submitted Paper - Neural networks and low-cost optical filters for plant segmentation PDF Print E-mail
Written by Ive Billiauws   
Monday, 12 April 2010 16:41

On March 29th, we have submitted the paper "Neural networks and low-cost optical filters for plant segmentation" for the IADIS Computer Graphics, Visualization, Computer Vision and Image Processing (CGVCVIP) 2010 conference. The 14th of June we'll know if our paper is accepted.



In this paper, we describe a low-cost system for image segmentation in agricultural applications. In applications such as weed detection and classification, plants must be segmented from the image. Our novel approach combines low-cost NIR filters  used on a standard camera with neural networks to achieve a significantly higher accuracy as compared to classic threshold techniques.

Last Updated on Monday, 12 April 2010 16:44
Seminar "Embedded Systems in precision-agriculture" PDF Print E-mail
Written by Ive Billiauws   
Wednesday, 18 November 2009 14:25

Tuesday 17/10/2009 we attended a seminar about embedded systems in precision-agriculture, organised by DSP Valley, ILVO and Innovatiesteunpunt voor Land- en Tuinbouw. The main reason for this seminar was bringing together both farmers and technologists/researchers for discussing various topics about electronics in agriculture.


The day started off with three general presentations to inform about current developments in electronics for agriculture: "Planting potatoes using a GPS" (by Septentrio), "Vision Techniques in Agriculture" (by IMEC) and "Reliability and Robustness of Electronics" (by TBP Electronics). This was followed by a discussion workshop. This allowed farmers to express their expectations about electronical aid in the future. This was very useful for us, as we're still looking for a user-case to develop during our project. In this workshop, we've also presented our SIVOL project.


We also exhibited a small demo of our project in the main hall. It consisted of some fresh weed, a USB webcam, a beagleboard and a display.  For this demo we've used OpenCV to write a C program for the beagleboard, which made a segmentation between weed and background and drew the edges around the extracted weed. Although this was a very simple demo, we've attracted the attention of several people who were interested in what we're doing.


This day was very interesting and useful for us as it led to a few new ideas and contacts. We'll be contacting them soon to discuss further cooperations.

Last Updated on Wednesday, 06 January 2010 17:26
SIVOL Website Online PDF Print E-mail
Written by Ive Billiauws   
Thursday, 15 October 2009 14:22

SIVOL is a project of the research cell EAVISE which is part of research group EmSD (Embedded Systems Design), at Campus De Nayer of Lessius University College. We recently started a new project: SIVOL or "Snelle Implementatie van Objectgebaseerde Objectherkenning in Landbouwtoepassingen".

SIVOL is a project in which we want to present a fast implementation of shape-based object recognition to be used for agricultural purposes. Current methods are based on pixel based color detection. This comes with a lot of disadvantages, such as blending of objects in the background. Therefore we want to explore the feasibility of more advanced image recognition methods for these applications.

The main reason why shape-based recognition isn't used for agriculture, are the need of a high processing speed and the higher requirements because of the environmental conditions. Dust, vibrations and illumination are only a few of the factors that can influence the recognition. Also the high variety of to-be-detected shapes requires more sophisticated algorithms.

On this website we'll soon publish information about the latest news, developments, events, ...

Come back soon to read more!



The SIVOL Team

Last Updated on Monday, 12 April 2010 16:43
SIVOL, Powered by Joomla!; Joomla templates by SG web hosting