An intelligent VegeCareAI tool for next generation plant growth management
Introduction
Recently, Artificial Intelligence (AI) based agriculture systems have attracted attention because Japan’s farming population has steadily decreased with agricultural worker’s average age has risen. Thus reducing the cultivated area where crops are harvested on a regular cycle. On the other hand, people are more concerned for safer and cost efficient food due to the impact of the COVID-19 epidemic and radioactive pollution [1], [2]. They have massively impacted every part of life, both economic and non-economic, including the agriculture sector. Likewise, most of the food produced for the food industry came from agricultural commodity products. A key role for solving the problem in fostering agricultural innovation is a collaboration between non-agriculture and agricultural technologies, such as edge computing, cloud computing and AI [3], [4], [5], [6], [7]. Cloud computing services can be used for big data analysis and massive amounts of batch jobs. However, edge computing is better for agriculture compared with cloud computing considering unfriendly conditions and remote locations of farms that may present network connectivity and bandwidth concerns. The purpose of edge computing is to provide a suitable, cost-effective alternative. In this way, edge-based agriculture systems are expected to improve agricultural products support to obtain high quality and safely to solve the shortage of agricultural workers [8], [9].
In order to improve the crops’ productivity, we proposed a classification tool considering tomato diseases [10]. We analyzed the classification performance in 6 classes of tomato diseases. The proposed tool predicted tomato mosaic virus correctly, but the testing results of septoria leaf spot is not high because the leaf color and leaf shape are similar.
In this paper, we propose a VegeCareAI tool for plant growth management considering vegetable classification, disease classification and insect pest classification. We use three kinds of classification functions to provide many guidelines for next generation agricultural worker. Our agricultural support system is focusing on edge AI, where daily learning is done in the cloud server and real-time prediction is done at the edge device.
The structure of the paper is as follows. In Sect. 2, we describe the related work. In Sect. 3, we present the proposed agricultural support system. In Sect. 4, we show the evaluation setting. In Sect. 5, we present the experimental results. Finally, conclusions and future work are given in Sect. 6.
Section snippets
Related work
During several years, some advanced systems are used for rice cultivation, which automatically can measure the water temperature and water level of the paddy field. In addition, a cultivation support system was proposed to collect the skill and knowledge of agricultural workers. For example, agricultural workers analyze or control the collected data with their mobile or tablet device. In this way, they can reduce the time needed for water management. Cloud computing can be used to analyze a
Overview of the proposed system
The structure of our proposed system is shown in Fig. 1.
The proposed system consists of the VegeCareAI tool used at the edge and the VegeCareAI system on the cloud. The VegeCareAI tool is a mobile application of the Android terminal that predicts the object and manages the plant growth for agricultural workers. We create an environment with edge devices in the field that can collect data via wireless communication and automatically upload it to the cloud. The cloud modules allow us to store
Training overall model
In Fig. 6 is shown the design of the proposed training model, which consists of four convolutional layers, four pooling layers and four fully connected layers based on the Keras sequential model. We use Rectified Linear Unit (ReLU) as the activation function to improve the representation of the model. Our CNN model considers the dropout layer to prevent the model from over-fitting. The output layer uses the softmax as an activation function to split the final result into multiple outcomes.
In
Vegetable classification results
The dataset contains 519 images for training, 58 images for validation and 58 images for testing. The learning rate for each epoch is more than 80%. The vegetable classification results are shown in Table 3. The accuracy results are not good for 100 epochs, especially for potato because their images are combined with the soil and their leaf looks similar to sweet potato. For 200 epochs, the accuracy is improved. The accuracy of the shiso reached 100%. In addition, the accuracy of the potato is
Conclusions
In this paper, we proposed a VegeCareAI for plant growth management considering three kinds of classification functions to provide guidelines for next generation agricultural workers. From the evaluation, we found that the following results: For vegetable classification, our training data for 300 epochs predicted six kinds of vegetables correctly. The accuracy of our model reached 100%. For disease classification, our training data for 400 epochs predicted more than 96% for potato leaves. For
Declaration of Competing Interest
None.
References (23)
- et al.
Plant leaf diseases detection and auto-medicine
Internet of Things
(2018) - et al.
Identification of plant leaf diseases using a nine-layer deep convolutional neural network
Computers and Electrical Engineering
(2019) - et al.
Automatic classification for field crop insects via multiple-task sparse representation and multiple-kernel learning
Computers and Electronics in Agriculture
(2015) COVID-19, agriculture, and food security in indonesia
Reviews in Agricultural Science
(2020)Agricultural aspects of radiocontamination induced by the fukushima nuclear accident - a survey of studies by the univ. of tokyo agricultural dept. (2011-2016)
Proceedings of the Japan Academy. Series B, Physical and Biological Sciences
(2018)- et al.
Internet of Things (IoT) and agricultural unmanned aerial vehicles (UAVs) in smart farming: A comprehensive review
Internet of Things
(2020) - et al.
Plant leaf disease detection and classification based on CNN with LVQ algorithm
Proceedings of the 3rd International Conference on Computer Science and Engineering (UBMK-2018)
(2018) - et al.
Genetic algorithm based internet of precision agricultural things (IopaT) for agriculture 4.0
Internet of Things
(2020) - et al.
Smart farming: Agriculture’s shift from a labor intensive to technology native industry
Internet of Things
(2020) - et al.
An overview of internet of things (IoT) and data analytics in agriculture: Benefits and challenges
IEEE Internet of Things Journal
(2018)
Internet of things (IoT) for smart precision agriculture and farming in rural areas
IEEE Internet of Things Journal
Cited by (2)
Unleashing the value of artificial intelligence in the agri-food sector: where are we?
2023, British Food JournalArtificial Intelligence: The Future of Sustainable Agriculture? A Research Agenda
2023, Proceedings of the Annual Hawaii International Conference on System Sciences