Research Center for Agricultural Information Technology, NARO

Launch of AI-based pest image diagnostic system via WAGRI

Towards popularizing AI pest diagnosis through service providers

NARO, Hosei university, and Northern System Service will start providing AI based pest image diagnostic system via WAGRI intended for agricultural information service providers. Each service provider can build and provide the pest image diagnosis service for general users by utilizing this system. This system accumulates and utilizes images sent from general users thereby it aims for the continuous improvement of diagnostic accuracy. As the first step, leaf surface disease discriminator was published for four types of vegetables & fruits such as tomato, cucumber, strawberry, and eggplant.


Overview

NARO, Hosei university, and Northern System Service will start providing AI based pest image diagnostic system that can be used by private agricultural information service providers through WAGRI. This system enables each service provider to build pest image diagnostic service and provide the same to general users such as producers.

To cope up with the decrease in skilled workers due to aging, inexperienced new farmers and new comers, and also as a countermeasure against the outbreak of pests in areas where there was no past outbreak due to global warming, this system is expected to be used for rapid diagnostic services at site. Moreover, by accumulating and utilizing images sent by the general users for pest diagnosis, this system aims to improve the accuracy of image discrimination by continuously improving AI.

A system based on AI leaf surface disease image diagnostic technology developed by the AI Pest Diagnosis Consortium (Representative Organization: NARO) for four types of vegetables & fruits such as tomato, cucumber, strawberry, and eggplant (where the diagnostic accuracy is 72.6% to 89.2%, which is high enough for actual operation) has been published in advance. AI based pest image diagnosis (diagnosis accuracy 82.1% to 85.4%) for the other parts such as leaf underside of the same target crops are scheduled to be released in April 2021. Simultaneously, for the private service providers who are considering the utilization of this system, free trial use of the demo app that realizes diagnostic services for farmers will be provided.

In order to promote data-driven agriculture utilizing WAGRI, even in the future, this system will deal with a wide range of pests and diseases of various agricultural products, by adding discriminator and disclosing it to private service providers through WAGRI, it is expected that the service can be quickly deployed to agricultural sites with low installation cost.


For Inquiries

Contact: http://www.naro.affrc.go.jp/english/inquiry/index.html

Centers・Institutes