Active learning with MaskAL reduces annotation effort for training Mask R-CNN on a broccoli dataset with visually similar classes

Article Environmental Sciences

Assessment of Different Object Detectors for the Maturity Level Classification of Broccoli Crops Using UAV Imagery

Vasilis Psiroukis et al.

Summary: The aim of this study was to automate the maturity classification of broccoli using object detection architectures and UAV images, and evaluate their effectiveness. The results showed that the automated approach performed well, especially with the best-performing architectures.

REMOTE SENSING (2022)

添加到收藏夹

Article Robotics

The effect of data augmentation and network simplification on the image-based detection of broccoli heads with Mask R-CNN

Pieter M. Blok et al.

Summary: The study aims to develop a robot capable of detecting broccoli heads using a deep learning algorithm to reduce labor costs. Data augmentation was found to improve the algorithm's generalization performance, with geometric transformations proving more effective than photometric transformations. The algorithm achieved successful generalization after including 5% of images from a specific cultivar in the training dataset.

JOURNAL OF FIELD ROBOTICS (2021)

添加到收藏夹

Article Agricultural Engineering

Image-based size estimation of broccoli heads under varying degrees of occlusion

Pieter M. Blok et al.

Summary: This study aimed to improve the estimation accuracy of broccoli head size by using deep learning algorithms to deal with occlusions. The ORCNN method outperformed the Mask R-CNN method in sizing occluded broccoli heads, showing better sizing performance.

BIOSYSTEMS ENGINEERING (2021)

添加到收藏夹

Article Agriculture, Multidisciplinary

Towards selective and automatic harvesting of broccoli for agri-food industry

Antonio Garcia-Manso et al.

Summary: This paper presents an approach to a vision system for automatically harvesting broccoli heads, based on deep learning techniques. The system is able to locate and classify broccoli heads in real-time digital images, achieving a classification accuracy of up to 97% in test conditions.

COMPUTERS AND ELECTRONICS IN AGRICULTURE (2021)

添加到收藏夹

Article Robotics