Machine Learning: The Potential To Help Protect The Public And The Environment

By: Matthew Kim

Project By: Kyle Kim and Matthew Kim

This paper references the science project, “Using machine learning to help protect the public from poison ivy, oak, and sumac and protect the ecosystem from invasive plants.”

Poison ivy, poison oak, and poison sumac may be harmful to humans. These plants release an oily resin, which can cause an allergic reaction, leading to rashes, blisters, and itchiness when in contact with the skin. Unfortunately, the public may not be aware of these poisonous plants and come in contact with them by accident.

Moreover, invasive species, such as English ivy and Japanese stilt grass, threaten our environment. These plants are aggressive towards native plants and compete with native plants for space to grow, water, sunlight, etc. Similar to poison ivy, poison oak, and poison sumac, the public may not be able to identify invasive plants to help mitigate the spread of these invasive plants.

Through the use of machine learning to develop a convolution neural network (CNN) model, my teammate and I were able to identify poison ivy, poison oak, poison sumac, and invasive plants, such as English ivy and Japanese stilt grass, via mobile devices. Doing so promotes early identification of these plants to prevent human exposure and help remove invasive species. For this project, two-thousand images of plants were used to train and validate the CNN models, and the CNN model was connected via web service using the Flask framework. An iOS prototype App was created to upload the plant images to the CNN model so that the CNN model can identify the plant and send the result back to the iOS app via an iPhone. The results revealed that the trained CNN models identified the plant images correctly with an accuracy rate of 98.6%.

Machine learning has the potential to help protect the public from poisonous plants and to help protect the ecosystem from invasive plants. Perhaps, in the future, machine learning could be used to map where invasive species or poisonous plants are located. The more work that goes into this machine learning will lead to more accurate results, thus protecting humanity and the environment.

REFERENCES

• https://www.tensorflow.org/hub/tutorials/tf2_image_retraining?hl=el

• https://towardsdatascience.com/creating-restful-apis-using-flask-and-python-655bad51b24

• http://www.connectionnewspapers.com/news/2016/mar/30/potomac-meet-marylands-invasive-plant-species/

• https://www.tensorflow.org/tutorials/images/classification

• https://extension.umd.edu/resource/introduction-invasive-plants

• https://www.analyticsvidhya.com/blog/2021/05/convolutional-neural-networks-cnn/

• https://www.mayoclinic.org/diseases-conditions/poison-ivy/symptoms-causes/syc-20376485

• https://www.poison.org/articles/poison-ivy-oak-and-sumac

• https://www.americanskin.org/resource/poisonivy.php

• https://www.flatworldsolutions.com/data-science/articles/7-applications-of-convolutional-neural-networks.php

• https://programminghistorian.org/en/lessons/creating-apis-with-python-and-flask

• https://www.nature.org/en-us/about-us/where-we-work/united-states/maryland-dc/stories-in-maryland-dc/maryland-invasive-species-taking-on-the-invaders-of-maryland/

• https://www.tensorflow.org/