Lawn Weed Detector

• Author: Chiyuan Cheng (cyuancheng AT gmail DOT com)
• Last updated: Aug 29, 2019

Table of Contents

1. Objective
2. Data
3. Result
4. License

Objective

To create an AI detector for common lawn weeds using computer vision and deep learning. The 7 common lawn weeds are:

• Annual Bluegrass
• Crabgrass
• Creeping Charlie
• Dandelion
• Dollar weed
• Oxalis
• White clover

Data

The training and test imaging dataset was collected from Google image and stored on google drive for analysis.

Result

• Image augmentation was used to improve the performance of deep learning in imaging classification.
• Dataset contains 1830 training data and 457 testing data.
• resnet50 was used for the deep learning model.
• 95% accuracy can be achieved to predict the lawn weed.
• The code for this project can be found here.

License

The credit was given fast.ai courses for the fast.ai package.