Wheat is one of the major crops in the world, with a global demand expected to reach 850 million tons by 2050 that is clearly outpacing current supply. The continual pressure to sustain wheat yield due to the world’s growing population under fluctuating climate conditions requires breeders to increase yield and yield stability across environments. We are working to integrate deep learning into field-based phenotypic analysis to assist breeders in this endeavour. We have utilised wheat images collected by distributed CropQuant phenotyping workstations deployed for multiyear field experiments of UK bread wheat varieties. Based on these image series, we have developed a deep-learning based analysis pipeline to segment spike regions from complicated backgrounds. As a first step towards robust measurement of key yield traits in the field, we present a promising approach that employ Fully Convolutional Network (FCN) to perform semantic segmentation of images to segment wheat spike regions. We also demonstrate the benefits of transfer learning through the use of parameters obtained from other image datasets. We found that the FCN architecture had achieved a Mean classification Accuracy (MA) >82% on validation data and >76% on test data and Mean Intersection over Union value (MIoU) >73% on validation data and and >64% on test datasets. Through this phenomics research, we trust our attempt is likely to form a sound foundation for extracting key yield-related traits such as spikes per unit area and spikelet number per spike, which can be used to assist yield-focused wheat breeding objectives in near future.

Currently, crop management through automatic monitoring is growing momentum, but presents various challenges. One key challenge is to quantify yield traits from images captured automatically. Wheat is one of the three major crops in the world with a total demand expected to exceed 850 million tons by 2050. In this paper we attempt estimation of wheat spikelets from high-definition RGB infield images using a fully convolutional model. We propose also the use of transfer learning and segmentation to improve the model. We report cross validated Mean Absolute Error (MAE) and Mean Square Error (MSE) of 53.0, 71.2 respectively on 15 real field images. We produce visualisations which show the good fit of our model to the task. We also concluded that both transfer learning and segmentation lead to a very positive impact for CNN-based models, reducing error by up to 89%, when extracting key traits such as wheat spikelet counts.