Modelling occlusions to improve wheat head detections

Monitoring the characteristics of the wheat heads regularly will assist the breeders to grow better varieties and the farmers with better management decisions. However, this task is still done manually which is a time consuming and can be erroneous in some scenarios. When the popular deep learning based object detectors (Faster-RCNN, Yolo) are used directly as it is, they tend to miss detecting the wheat heads that are occluded by leaves which impacts the accuracy of yield prediction and is not desirable. The students will be using baseline pytorch code (FasterRCNN) to study the influence of random masking based approaches for modelling leaf occlusions and further improve the wheat head detection accuracy. 


Interested students can contact me at

Apply Online

Explain why you want this job and what qualifies you specifically (max 3000 characters)
In here, include details about your education to date (modules taken and grades achieved) as well as any other relevant work experience
If you have a professional online portfolio or profile, please link it here

Fields with (*) are compulsory.