Sugar Cane Grading from Photos Using Convolutional Neural Networks
| Autor: | Sally E. Goldin, Phuong Pham |
|---|---|
| Rok vydání: | 2019 |
| Předmět: |
biology
business.industry Computer science Sugar cane Sugar industry Feature extraction Growing season 020206 networking & telecommunications 02 engineering and technology biology.organism_classification Machine learning computer.software_genre Convolutional neural network 0202 electrical engineering electronic engineering information engineering 020201 artificial intelligence & image processing Artificial intelligence Cane business Grading (education) Sugar yield Sugar computer |
| Zdroj: | 2019 4th International Conference on Information Technology (InCIT). |
| Popis: | Sugar manufacturing companies need up-to-date information about sugar cane condition in order to forecast sugar yield. However, the wide spatial distribution of cane fields in Thailand makes exhaustive surveys of sugar cane condition impractical. In this paper, we report on initial efforts to apply convolutional neural networks (CNN) to classify sugar cane condition based on ground level photographs of sugar cane fields at different stages of the growing season. Our long term goal is to create a system where farmers can submit mobile phone photos of their fields that will be evaluated for cane quality using a machine learning (ML) model. Results suggest that this approach has promise, but that we need a larger set of exemplars to create a model that can provide classification performance accurate enough for operational use. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|