UTFacultiesEEMCSDisciplines & departmentsPSEducationAssignment: Optimization of seedling detection by using instance segmentation

Assignment: Optimization of seedling detection by using instance segmentation

Optimization of seedling detection by using instance segmentation


Problem statement

Track32 deploy and further develop a service for fully automated seedling measurements in collaboration with Corvus Drones. For a grower, it is important to have an objective analysis of his seedlings to reliably assess the sellability and the status. Currently, the seedlings are assessed manually by several employees, which results in subjectivity between trays for the same employee and subjectivity between employees for the same tray. Hence, a grower needs to includes a margin for selling his stock, and is selling less than his production. Our seedling measurement service could serve that problem by delivering objective analysis. However, this requires a good performing seedling detection.

Task

The task for this project is to investigate which adaptations into the model architecture (for example yolact vs mask-rcnn), augmentations and training adaptations (dynamic learning rate, validation metrics) are promising for a high performing seedling detection software module. Furthermore, we could also look into publicly available seedling datasets which may be useful for increasing the model performance. You will perform several experiments and in the end you will present an overview of the results including your thoughts about the experiment with the highest model performance. Your ideas will be integrated into our current train module, to acquire the highest performance as possible.   

Work

20% Theory, 60% implementation, 20% writing report

Contact:

Le Viet Duc, v.d.le@utwente.nl

About Track32:

At Track32 we produce innovative computer vision and AI software. Making the technology easily accessible, so that it becomes part of your organization’s natural intelligence.

Track32 provides state-of-the-art computer vision and AI algorithms, and we integrate them into existing or new hardware and software systems. We are experts in processing all types of visual and non-visual data, using deep learning and other methods. Track32 excels at analyzing the user’s requirements and matching them with technically feasible and cost-effective solutions. Using our computer vision and AI solutions leads to huge cost savings and massively increased operational efficiency for our customers.

Computer vision and AI are generic technologies that can be used in any application domain. We serve a wide spectrum of customers in the agricultural supply chain, but also players in other markets such as post harvest, material handling, spatial planning, healthcare and the life sciences. We serve commercial companies as well as research institutes and government.