Pieter Blok Wins EMVA Young Professional Award

EMVA has awarded its 2023 Young Professional Award to Pieter Blok, a researcher of deep learning and computer vision at Wageningen University and Research in the Netherlands. Blok earns the award for his work “advanced deep learning for harvest robotics.”

The award was announced during the 21st EMVA Business Conference. Blok, who presented his work as part of the regular conference program, holds a bachelor’s degree and a master’s degree in agricultural engineering from Wageningen.

The goal of Blok’s Ph.D. thesis was to research and develop new machine vision models that can help a harvesting robot to deal with variations and uncertainties. The focus of the thesis was on three tasks that must be performed by every harvesting robot: crop detection, crop size estimation, and crop quality determination.
Pieter Blok, left, and EMVA President Chris Yates. Courtesy of EMVA.

Pieter Blok (left) and EMVA president Chris Yates. Courtesy of EMVA.

In the context of crop detection, Blok’s thesis focused on improving the generalization performance of convolutional neural networks (CNNs) to deal with variations within the same crop. Usually, there are many variations within a crop, which can create problems with the CNN’s ability to generalize. Blok’s research focused on applying different types of data augmentation to optimize the training of a CNN. With geometric data augmentation (rotation, cropping, and scaling of the image), the CNN was able to better generalize to multiple crop varieties.

In size estimation, a harvesting robot determines whether a crop should be harvested or left in the field to grow further. Challenges can arise, for example, when a crop is occluded by leaves, which reduces visibility. Blok addressed this with amodal perception, which is the ability to predict both the visible and occluded parts of objects in an image. By integrating this into a CNN, the larger amodal shape of occluded crops could be accurately estimated for 3D crop size estimation and robot positioning.

The third aspect of Blok’s thesis, quality determination, focused on using active learning to automatically select and annotate sporadically occurring plant diseases. The newly developed active learning method used the output of the CNN to select unlabeled images about which the network was most uncertain. These selected images were then interactively labeled in a semi-supervised way and used to retrain the network. This active learning method significantly reduced the annotation effort by 1400 image annotations (120 annotation hours).

Beginning July 1, Blok will be an assistant professor at the Laboratory of Field Phenomics at the University of Tokyo. There, he will focus on machine learning and image processing technologies for plant phenotyping.

Published: May 2023

Glossary

convolutional neural network: A powerful and flexible machine-learning approach that can be used in machine vision to help solve difficult problems. Inspired by biological processes, multiple layers of neurons process portions of an image to arrive at a classification model. The network of neurons is trained by a set of input images and the output classification (e.g., picture A is of a dog, picture B is of a cat, etc.) and the algorithm trains the neuron connection weights to arrive close to the desired classification. At...
machine vision: Machine vision, also known as computer vision or computer sight, refers to the technology that enables machines, typically computers, to interpret and understand visual information from the world, much like the human visual system. It involves the development and application of algorithms and systems that allow machines to acquire, process, analyze, and make decisions based on visual data. Key aspects of machine vision include: Image acquisition: Machine vision systems use various...
computer vision: Computer vision enables computers to interpret and make decisions based on visual data, such as images and videos. It involves the development of algorithms, techniques, and systems that enable machines to gain an understanding of the visual world, similar to how humans perceive and interpret visual information. Key aspects and tasks within computer vision include: Image recognition: Identifying and categorizing objects, scenes, or patterns within images. This involves training...
deep learning: Deep learning is a subset of machine learning that involves the use of artificial neural networks to model and solve complex problems. The term "deep" in deep learning refers to the use of deep neural networks, which are neural networks with multiple layers (deep architectures). These networks, often called deep neural networks or deep neural architectures, have the ability to automatically learn hierarchical representations of data. Key concepts and components of deep learning include: ...

Browse Cameras & Imaging, Lasers, Optical Components, Test & Measurement, and more.