- peng.yue@sparksoft.uk
- +44 (0) 7780 856 888
- Willesden Green, London, UK, NW2
Adaptive AI-enabled Mobile Intelligence Solutions for Climate-smart Pest Management
Our Team
SparkSoft Limited has 20 years experiences in software architecture and development and has been working with a number of international clients. and we are glad to work with Terravost Limited, which is one of the UK leading agribusiness managment consultancy, and OAK group at the computer science department of University of Sheffield (UoS) which is proficient in machine learning based model such as PestNet and Human Activity Recognisation (HAR) on this AI-enabled mobile intelligence system. With these complementary backgrounds and expertise of all three partners, this collaboration demonstrates excellent results.
Our Goal
We investigated the technical feasibility of integrating contextual and visual information with adaptive AI technique and aim to a mobile solution that offers:
- rapid detection and quantification of crop pests by mobile devices;
- efficient forecasting of accepted pest thresholds for climate-smart management;
Our Dataset
The PhD students collects the crops picture with the Terravost Limited crop management expert in Yorkshire.
Due to weather condition in winter of 2022, we were not able to collect all the dataset, however, we were able to collect pest and crop pictures dataset by visiting the farms in Yorkshire from Nov 2022 to Jan 2023. Over 1,000 images were collected and added to our dataset during these 3 months. We also collected 19 publicly or privately available datasets related to crop pests and diseases, including 291,185 images for pest classification tasks, disease classification and other tasks. Last but not least, we were able to collect pest images through search engines. 230,000 images (without commercial licenses) and 4,000 images (with commercial licenses) were collected and were required to further data cleaning and annotation.
Our Progress
Our current deliverables include:
- A demonstration of mobile intelligence solution with advanced AI models for rapid detection and effective measure of multiple crop pests (MCP) in cereal, rapeseed, and potato cropping systems;
- An evaluation report resulting from the above techniques that can be used to place pest quantities into the contexts of thresholds regional pesticide resistance status to advise growers on management options.
