Multi-Scale Context Aggregation for Strawberry Fruit Recognition and Disease Phenotyping

Abstract

Timely harvesting and disease identification of strawberry fruits is a major concern for commercial level cultivators. Failing to harvest the grown strawberries can result in the fruit rotting which makes their damaged tissues more prone to grey mold pathogens. Immediate removal of the overgrown or diseased strawberries is inevitable to curb the mass spreading of the pathogen. In this paper, we propose a deep learning-based framework to identify three different strawberry fruit classes (unripe, partially ripe and ripe), as well as a class of overgrown or diseased strawberries. We equip the proposed convolutional encoderdecoder network with three different modules. One for adaptively controlling receptive filed size of the network to detect objects of multiple sizes. Second for controlling the flow of salient features (information) to the deeper layers of the network and the other for controlling the architecture's computational complexity. These modules combined, outperform the previous state-of-the-art semantic segmentation networks on the task of strawberry fruit phenotyping. We also introduce a dataset collected from different farms to evaluate the performance of the network. Quantitative and qualitative results show that notwithstanding heterogeneity in the data and the effect of the real-field variations, our approach produced remarkable results with a 3% increase in mean intersection over union as compared to the other state-of-the-art networks and was able to recognize diseased fruits with a precision of 92.45%.

High‐throughput phenotyping for breeding targets—Current status and future directions of strawberry trait automation

“…The variation in environmental conditions found in real‐world agricultural settings poses a further challenge to image‐based phenotyping and has been highlighted by several authors in the context of strawberry phenotyping. Illumination changes, due to variation with the time of day, weather and shadow (Heylen et al., 2021; Ilyas et al., 2021; Kirk et al., 2020; Lin & Chen, 2018; Yu et al., 2019; Zhou et al., 2020) pose one such impediment. However, CNNs are reasonably robust to variance in illumination, and this can be further addressed through methods such as merging features from different colour spaces (Kirk et al., 2020).…”

“…As fruit and flower counts are also vital for yield forecasting applications, the precise detection and counting of these components have been a research focus for high‐throughput methods, with around half of the literature pertaining to automation of phenotypic traits centring around this. In image‐based phenotyping, CNNs have been used to address fruit (Chen et al., 2019; Fan et al., 2022; Ilyas et al., 2021; Kerfs et al., 2017; Kim et al., 2020; Kirk et al., 2020; Lamb & Chuah, 2018; Yu et al., 2019; Zhang et al., 2022; Zhou et al., 2020) and flower (Heylen et al., 2021; Lin & Chen, 2018) detection in real‐world agricultural conditions, as they offer greater robustness to the varying environmental conditions experienced than traditional machine learning methods that use manually defined features. However, a challenge for detection in real‐world environments is occlusion, which can be minimised by selecting an appropriate viewpoint from which to collect data so as to maximise the prominence of the organ of interest in the image.…”

Thangka Image Segmentation Method Based on Enhanced Receptive Field

“…Recently, DAM [14] proposed a dense attention module enabling hierarchical adaptive feature fusion by exploiting interchannel and intra-channel relationships. The research introduces the adaptive receptive field and channel selection modules that enable the network to tackle variable-sized instances and correlated feature maps [15].…”

Accurate segmentation of green fruit based on optimized mask RCNN application in complex orchard

“…In theory, it can be mapped to any function to solve more complex problems (Hussain et al, 2020 ; Naranjo-Torres et al, 2020 ; Jia et al, 2021 ; Lin et al, 2021a ). Ilyas proposed convolutional encoder–decoder network for strawberry fruit maturity recognition and diseased fruit and introduced adaptive receptive field, channel selection module, and bottleneck module to realize the accurate recognition of strawberry fruit, but the model could not segment a single target (Ilyas et al, 2021 ). Aiming at many problems in complex orchard environment, Kang proposed a one-stage detector DaSNet-v2.…”