The Australian dryland grain-cropping landscape occupies 60 Mha. The broader agricultural sector (farmers and agronomic advisors, grain handlers, commodity forecasters, input suppliers, insurance providers) required information at many spatial and temporal scales. Temporal scales included hindcasts, nowcasts and forecasts, at spatial scales ranging from sub-field to the continent. International crop-monitoring systems could not service the need of local industry for digital information on crop production estimates. Therefore, we combined a broad suite of satellite-based crop-mapping, crop-modelling and data-delivery techniques to create an integrated analytics system (Graincast™) that covers the Australian cropping landscape. In parallel with technical developments, a set of user requirements was identified through a human-centred design process, resulting in an end-product that delivered a viable crop-monitoring service to industry. This integrated analytics solution can now produce crop information at scale and on demand and can deliver the output via an application programming interface. The technology was designed to underpin digital agriculture developments for Australia. End-users are now using crop-monitoring data for operational purposes, and we argue that a vertically integrated data supply chain is required to develop crop-monitoring technology further.

In Australia, digital agriculture is considered immature and its adoption ad hoc, despite a relatively advanced technology innovation sector. In this review, we focus on the technical, governance and social factors of digital adoption that have created a disconnect between technology development and the end user community (farmers and their advisors). Using examples that reflect both successes and barriers in Australian agriculture, we first explore the current enabling technologies and processes, and then we highlight some of the key socio-technical factors that explain why digital agriculture is immature and ad hoc. Pronounced issues include fragmentation of the innovation system (and digital tools), and a lack of enabling legislation and policy to support technology deployment. To overcome such issues and increase adoption, clear value propositions for change are necessary. These value propositions are influenced by the perceptions and aspirations of individuals, the delivery of digitally-enabled processes and the supporting legislative, policy and educational structures, better use/conversion of data generated through technology applications to knowledge for supporting decision making, and the suitability of the technology. Agronomists and early adopter farmers will play a significant role in closing the technology-end user gap, and will need support and training from technology service providers, government bodies and peer-networks. Ultimately, practice change will only be achieved through mutual understanding, ownership and trust. This will occur when farmers and their advisors are an integral part of the entire digital innovation system.

This paper reviews early experiences, expectations and obstacles concerning the adoption of digital technologies in Australian livestock systems. Using three case studies of publicly-available information on Australia’s red meat industry, we identify the process of digitally enhanced value creation according to four themes: (1) supply chain operability; (2) product quality; (3) animal welfare; and (4) innovation and learning. We find reasons for both optimism and pessimism concerning the adoption of digital agriculture. While digital technology is being offered by various stakeholders to support collaboration within supply chains, it is also being met with scepticism amongst some producers who are not actively engaging with a digital transformation. We identify that the ‘technology fallacy’, which proposes that organisations, people, learning and processes are as important to digital transformation as the technology itself; but while digital technologies enable change, it is the people who determine how quickly it can occur. We argue that – since quality appears to be the major basis on which Australian red meat producers will compete in global markets – the broad adoption of digital technology will prove increasingly essential to future growth and sustainability of this supply chain.

Digital agriculture (DA) can contribute solutions to meet an increase in healthy, nutritious, and affordable food demands in an efficient and sustainable way. South America (SA) is one of the main grain and protein producers in the world but the status of DA in the region is unknown. A systematic review and case studies from Brazil, Argentina, Uruguay, and Chile were conducted to address the following objectives: (1) quantify adoption of existing DA technologies, (2) identify limitations for DA adoption; and (3) summarise existing metrics to benchmark DA benefits. Level of DA adoption was led by Brazil and Argentina followed by Uruguay and at a slower rate, Chile. GPS guidance systems, mapping tools, mobile apps and remote sensing were the most adopted DA technologies in SA. The most reported limitations to adoption were technology cost, lack of training, limited number of companies providing services, and unclear benefits from DA. Across the case studies, there was no clear definition of DA. To mitigate some of these limitations, our findings suggest the need for a DA educational curriculum that can fulfill the demand for job skills such as data processing, analysis and interpretation. Regional efforts are needed to standardise these metrics. This will allow stakeholders to design targeted initiatives to promote DA towards sustainability of food production in the region.

This paper presents the way the digital transformation of the agricultural sector is implemented in Europe and in France. It describes the main European and national strategies, the structure of research and innovation initiatives, and the investment in capacity building to foster innovation, adoption and use. More specifically, the French research and innovation ecosystem on digital agriculture is described. The actors involved come from different organisations, such as research and higher educational institutes, government agencies, agricultural technology (AgTech) companies, farmer unions etc., and work together by means of associations (e.g. Robagri), networks (e.g. RMT Naexus, DigiFermes, Fermes Leader), or living labs (e.g. Occitanum) on both digital technology assessment and co-design. Additionally, support is devoted to capacity building (e.g. Le Mas numérique, Mobilab) and a better understanding of the drivers of adoption and use of digital technologies (e.g. FrOCDA). Among these various organisations, #DigitAg, the Digital Agriculture Convergence Lab, has been created to foster interdisciplinary research on digital agriculture. All these initiatives aim to use digital technologies to support the European Green Deal, Farm-to-Fork and Biodiversity strategies as well as the French orientation towards more agroecological practices for safer and more sustainable food systems. Even though this organisational ecosystem is developing fast, the objective of encouraging the coevolution of both digital and green transformations is not without challenges that still need to be overcome, either through new research, innovations, initiatives or collaborations between the actors involved.

Agriculture is central to the Indian economy and suffers from widespread operational inefficiencies that could be corrected by the use of digital agriculture technologies (DA). We review and synthesise available literature concerning digital agriculture in India and anticipate its transformative potential in the coming decade. Although the initial growth of DA was more conspicuous in the downstream sectors and high-value crops, reaching smallholder farmers upstream is slowly emerging despite significant obstacles such as small fragmented holdings, inadequate data infrastructure and public policy, and unequal access to digital infrastructure. Agri-tech enables innovation at many locations within value chains, and a steady shift is occurring in change from individual farms to the whole value chain. Technology in the sector is progressing from information and communication technology-based solutions to Internet of Things and artificial intelligence–machine learning-enabled services. India’s public policy shows signs of a longstanding investment and collaboration in the sector, with an explicit focus on data infrastructure development. We find smallholder predominance, diversity in production systems, the predominance of commodity crops, proximity to urban markets, and public policy as the major factors of DA’s success in India. A stocktake of the available technologies and their applications by the public sector, tech giants, information technology leaders and agri-food tech startups in India strongly indicates a digital transformation of Indian agriculture. However, given the federal structure of governance and agriculture being a state (province) subject, we need to wait to see how DA policies are rolled out and taken up across the country.

In the upcoming years, global changes in agricultural and environmental systems will require innovative approaches in crop research to ensure more efficient use of natural resources and food security. Cutting-edge technologies for precision agriculture are fundamental to improve in a non-invasive manner, the efficiency of detection of environmental parameters, and to assess complex traits in plants with high accuracy. The application of sensing devices and the implementation of strategies of artificial intelligence for the acquisition and management of high-dimensional data will play a key role to address the needs of next-generation agriculture and boosting breeding in crops. To that end, closing the gap with the knowledge from the other ‘omics’ sciences is the primary objective to relieve the bottleneck that still hinders the potential of thousands of accessions existing for each crop. Although it is an emerging discipline, phenomics does not rely only on technological advances but embraces several other scientific fields including biology, statistics and bioinformatics. Therefore, establishing synergies among research groups and transnational efforts able to facilitate access to new computational methodologies and related information to the community, are needed. In this review, we illustrate the main concepts of plant phenotyping along with sensing devices and mechanisms underpinning imaging analysis in both controlled environments and open fields. We then describe the role of artificial intelligence and machine learning for data analysis and their implication for next-generation breeding, highlighting the ongoing efforts toward big-data management.

ContextInsects are a major threat to crop production. They can infect, damage, and reduce agricultural yields. Accurate and fast detection of insects will help insect control. From a computer algorithm point of view, insect detection from imagery is a tiny object detection problem. Handling detection of tiny objects in large datasets is challenging due to small resolution of the insects in an image, and other nuisances such as occlusion, noise, and lack of features.

AimsOur aim was to achieve a high-performance agricultural insect detector using an enhanced artificial intelligence machine learning technique.

MethodsWe used a YOLOv3 network-based framework, which is a high performing and computationally fast object detector. We further improved the original feature pyramidal network of YOLOv3 by integrating an adaptive feature fusion module. For training the network, we first applied data augmentation techniques to regularise the dataset. Then, we trained the network using the adaptive features and optimised the hyper-parameters. Finally, we tested the proposed network on a subset dataset of the multi-class insect pest dataset Pest24, which contains 25 878 images.

Key resultsWe achieved an accuracy of 72.10%, which is superior to existing techniques, while achieving a fast detection rate of 63.8 images per second.

ConclusionsWe compared the results with several object detection models regarding detection accuracy and processing speed. The proposed method achieved superior performance both in terms of accuracy and computational speed.

ImplicationsThe proposed method demonstrates that machine learning networks can provide a foundation for developing real-time systems that can help better pest control to reduce crop damage.

Context. Most weed species can adversely impact agricultural productivity by competing for nutrients required by high-value crops. Manual weeding is not practical for large cropping areas. Many studies have been undertaken to develop automatic weed management systems for agricultural crops. In this process, one of the major tasks is to recognise the weeds from images. However, weed recognition is a challenging task. It is because weed and crop plants can be similar in colour, texture and shape which can be exacerbated further by the imaging conditions, geographic or weather conditions when the images are recorded. Advanced machine learning techniques can be used to recognise weeds from imagery.

Aims. In this paper, we have investigated five state-of-the-art deep neural networks, namely VGG16, ResNet-50, Inception-V3, Inception-ResNet-v2 and MobileNetV2, and evaluated their performance for weed recognition.

Methods. We have used several experimental settings and multiple dataset combinations. In particular, we constructed a large weed-crop dataset by combining several smaller datasets, mitigating class imbalance by data augmentation, and using this dataset in benchmarking the deep neural networks. We investigated the use of transfer learning techniques by preserving the pre-trained weights for extracting the features and fine-tuning them using the images of crop and weed datasets.

Key results. We found that VGG16 performed better than others on small-scale datasets, while ResNet-50 performed better than other deep networks on the large combined dataset.

Conclusions. This research shows that data augmentation and fine tuning techniques improve the performance of deep learning models for classifying crop and weed images.

Implications. This research evaluates the performance of several deep learning models and offers directions for using the most appropriate models as well as highlights the need for a large scale benchmark weed dataset.