My first objection Detection model – Tensorflow

I trained this model by following a tutorial by a youtube creator called sentdex. You can find his videos here

https://www.youtube.com/user/sentdex

To train the model I took 150 photos of broad leaved docks in one of our grassland fields. I used a program called rectlabel for OSX to label all my docks in the image. This bounding box tells Tensorflow where to look for the weeds/docks. Tensorflow then trains itself to place accurate boxes around all the weeds in the data set by constructing and training a neural network.

 

2 thoughts on “My first objection Detection model – Tensorflow”

  1. Hey!!
    I also want to build project similar to your but I want cut grass(in short I am building lawn mower)
    So, I am not getting any proper resource to study how to do that so can you provide me resources that you used while developing our model for project.
    I want to create my own dataset of grass images train it.
    Please ,respond.

Leave a Reply to Gavin Cancel reply

Your email address will not be published. Required fields are marked *