A classic approach to detect potato diseases

Abstract : Nowadays, agriculture is mainly influenced by technology. Hence, computer vision is a crucial component to improve crops harvests. In this work, we propose an automatic solution to detect potato pests diseases. Our solution combines image processing and machine learning and finds that potatoes disease can be detected with high accuracy. We compare traditional classifiers and find that SVM is the best performer. We also analyze our results with PCA and we observe that many classes are linearly separable. This is an ongoing effort to improve crop harvest
Complete list of metadatas

Cited literature [12 references]  Display  Hide  Download

https://hal.archives-ouvertes.fr/hal-02225215
Contributor : Jeffri Murrugarra Llerena <>
Submitted on : Thursday, August 1, 2019 - 6:03:52 AM
Last modification on : Thursday, August 1, 2019 - 2:02:39 PM

Identifiers

  • HAL Id : hal-02225215, version 1

Citation

Jeffri Murrugara. A classic approach to detect potato diseases. "LatinX in AI Research at ICML 2019, Jun 2019, California, United States. ⟨hal-02225215⟩

Share

Metrics

Record views

36

Files downloads

31