Images become an increasingly more important source of information. Photos of plants, for instance, can be used for plant phenotyping and monitoring, satelite images, provide information about ground cover, and videos can be used to monitor the health of animals. Computer vision offers computational methods to process image and video data in order to extract relevant information.
In this 5-day course, you will learn about the basics of computer vision, from the acquisition of good quality images to the use of Python programming to implement computer-vision solutions to extract relevant information for your domain. You will learn about more traditional image-processing techniques as well as state-of-the-art deep neural networks to process images and videos. The first three days cover generic topics that are relevant in all application domains. We then provide some sessions of choice depending on your interest. At the final afternoon, we take time to discuss your individual projects.
The course is hands on. Every session will start with a brief theoretical introduction, followed by practical assignments where you learn how to apply the theory in practice using Python programming.
The course will deal with the following topics:
- Image acquisition and the representation of images and videos in a computer.
- Methods for image pre-processing, such as noise filtering, and the extraction of edges and contours.
- Introduction to deep learning with specific applications for image classification, segmentation and object detection.
- Image-based remote sensing to extract plant traits from UAV data (optional).
- Pose estimation to extract information about gait of humans/animals (optional).
- Working with spectral images to estimate plant content and diseases (optional).
- Tracking of multiple objects in videos to derive animal activity (optional).
Monday 5 October
- Getting to know each other
- Introduction to computer vision and images
Image acquisition
Tuesday 6 October
- Image pre-processing
- Edges and contours
Image segmentation and cocos
Wednesday 7 October
- Introduction to deep learning
- Deep learning for image classification
Deep learning for image segmentation
Thursday 8 October
- Deep learning for object detection and instance segmentation
Image-based remote sensing OR Animal/human pose estimation
Friday 9 October
- Working with spectral images OR Tracking of objects/animals in videos
Bring you own images
- Gert Kootstra - Agricultural Biosystems Engineering
- Gerrit Polder - Robotics and Automation
- David Rapado Rincon - Agricultural Biosystems Engineering
- Ricardo da Silva Torres - Artificial Intelligence
- Lammert Kooistra - Geo-information Science and Remote Sensing Laboratory
| Target Group | The course is aimed for PhD students, postdocs and staff members who are working or planning to work with images or videos for their research, but who have little knowledge about computer-vision methodologies and the underlying theories. |
| Prior knowledge | We assume some skills and experience in programming in Python. If you do not have that, you can contact the organizers in advance to get advice on how to acquire these skills using online resources. |
| Accommodation | Accommodation is not included in the course |
| Location | Wageningen University |
| Fees1 | FEE |
| PhD candidates of PE&RC/WIMEK/VLAG/WIAS with approved TSP and WU EngD candidates | € 355,- |
| PE&RC postdocs and staff | € 710,- |
| Other academic participants | € 750,- |
| Non-academic participants | € 1460,- |
1 The course fee includes lunch, coffee/tea, and a reader. It does not include accommodation.
Note:
PE&RC Cancellation Conditions
IMPORTANT: ALWAYS read the Cancellation conditions for PE&RC courses and activities.