Eurizon.eu

Organization: University of Southern Denmark

Location: Odense, Denmark

Field: computer science

Requirements:

The successful candidate will be enrolled at this university in accordance with faculty regulations and the Danish Ministerial Order on the PhD Program at the Universities (PhD order).

Students must have either graduated from a Scandinavian MSc program, an MSc program in English, have English as their native language, or alternatively have passed a language test at the level demanded by University of Southern Denmark, Faculty of Engineering (minimum scores: IELTS 6.5 points; Paper-based TOEFL 575 points; Computer based TOEFL 230, Internet-based TOEFL 88 points).

Abstract:

The Institute of Chemical Engineering, Biotechnology and Environmental Technology (KBM), Faculty of Engineering, University of Southern Denmark is seeking a highly motivated graduate to undertake research in vision based segmentation and temporal quantification in the spatial development of textured structures.

Description:

The position is in close cooperation with the Mærsk Mc-Kinney Møller institute (MMMI). The Ph.D. position is part of two larger projects: ⅔ funding - GUDP (Grønt- Udviklings- og DemonstrationsProgram) project with the title: “Graduation of fungicides and herbicides in potatoes and cereals”; and ⅓ funded by Mærsk Mc-Kinney Møller institute (MMMI) via a project titled “DiagnoseBot - Increasing the effectiveness of wound treatment and operations” (DiagnoseBot - Effectivisering af sårbehandling og operationer). Both projects will initially share active 3D vision systems for image acquisition of crop/weed scenes and human wounds. The PhD student will collaborate closely with the Cognitive Vision Lab (CoViL) research group at SDU which already has a powerful GPU accelerated framework for creating a vision system based on multiple view geometry referred to as Cognitive Vision Library (CoViS).

The focus will be on

● Robust segmentation based on 2D and 3D vision (using the CoVIS framework)
● Extraction of features for classification
● Robust quantification of the early development of textured structures
● Real-time implementation

The Ph.D. project must contribute to the research within image based segmentation and quantification of temporal-spatial development of textured structures of the following living case objects:

● In the early growth stages perform crop / weed ratio estimation in the open field environments of the farming industry.
● Skin damage on humans for example in the form of wounds.

It is believed that the development of sensor-based site-specific plant protection will be one of the key elements in achieving the politically stated goals of reducing the use of pesticides. Development of sensor-based site-specific crop protection systems are a prerequisite for agricultural industry realizing the potential of reducing pesticide use caused by variable presence of pests within the individual fields and between fields. Only through the use of automated sensor-based site-specific pest control can the reduction potential be achieved together with an improved operating economy. Thus the latter is a prerequisite for maintaining the competitiveness within the primary agricultural sector.
In cooperation KBM and MMMI has developed a very promising algorithm for robust quantification of the ratio between monocots and dicots in the early growth stages. The algorithm is independent of partial leaf occlusion. The algorithm has been developed within the high technology project: “Den Intelligente SprøjteBom” (DISB) (The intelligent Spraying boom). A patent has been applied for the ideas behind the algorithm. This PhD. must develop and validate the algorithm further.

The developed principles is sought to be expanded to another problem domain, the medico sector. In DiagnoseBot the system shall acquire images of damaged skin areas and track their temporal development.
It is the purpose of the project to create an ideal observer, which is able to express how a successful treatment of a case of skin damage should look. To track the temporal development of wounds. This challenge has multiple parallels to the vision challenges within crop / weed identification and temporal quantification.

Applicants for the PhD scholarship must have a masters degree within the subjects applied computer science, machine vision and/or statistics. The applicants must have experience with:

● Sensor technology
● Image analysis within the domain of crop/weed
● Feature extraction
● Embedded software and hardware design
● Acceleration of signal processing algorithms, using Graphical Processing Units (GPU)
 
Appointment as a PhD Research Fellow is for three years. Employment stops automatically at the end of the period. The holder of the Fellowship is not allowed to have other paid employment during the three-year period.

The successful applicant will be employed in accordance with the agreement of 1.October 2008 on salaried PhD scholars between the Ministry of Finance and AC (the Danish Confederation of Professional Associations).

Please send the application, marked Job ID 114006 and enclosures, which should include the following:

• Detailed CV, including list of publications (if any)
• Copy of diploma
• A brief one-page proposal for study plan

to University of Southern Denmark, The Faculty of Engineering, Campusvej 55, 5230 Odense M, Denmark

Deadline: 09-02-2011

Contacts:

Link: http://www.jobs.sdu.dk/vis_stilling.php?id=6346&lang=eng

Email: rasj@kbm.sdu.dk

Email: lpc@kbm.sdu.dk

If you apply for this position please say you saw it on eurizons.eu