25 points by machine_vision_enthusiast 6 months ago flag hide 29 comments
johnsmith 6 months ago next
Great project! I've been looking for something like this for my garden.
randomtechy 6 months ago prev next
Thanks for sharing! How accurate is the TensorFlow model in detecting the diseases?
johnsmith 6 months ago next
The model has an accuracy of around 95% on our test dataset.
gpugirl 6 months ago prev next
Impressive results! Did you use a pre-trained model or train it from scratch?
johnsmith 6 months ago next
We used a pre-trained TensorFlow model and fine-tuned it on our dataset.
piuser 6 months ago prev next
This is amazing! I would love to see the code on Github.
johnsmith 6 months ago next
We'll be sure to open-source the project and post an update here once it's ready!
edtechguy 6 months ago prev next
How many plants can you monitor in real-time?
johnsmith 6 months ago next
With our current setup, we can monitor up to 5 plants at a time. But we're working on scaling it up!
cleverusername 6 months ago prev next
How did you power the Raspberry Pi? External power bank or something else?
johnsmith 6 months ago next
We used a portable rechargeable power bank that supports USB devices. Works great so far!
mrrobot 6 months ago prev next
Do you have any plans for a future update with more disease classifications?
johnsmith 6 months ago next
Definitely! That's a great suggestion. We plan to expand the dataset to include more diseases.
hugenums 6 months ago prev next
This is such a cool use case of TensorFlow! Thanks for showing us your project!
johnsmith 6 months ago next
You're welcome! Glad you like it. We're excited to see the impact this will have on gardening and farming.
inquisitive 6 months ago prev next
What's the implementation process in detecting the plant diseases? I mean, how does it actually work step by step?
johnsmith 6 months ago next
Sure, we first capture real-time images or videos of the plant leaves. Then we process and preprocess the visual data, and lastly, feed it to TensorFlow to perform the inference and classification.
constantruntime 6 months ago prev next
What about false positives? How do you ensure the right disease is identified?
johnsmith 6 months ago next
Great question. We include multiple example images of each disease in our dataset and use a validation set to minimize false positives.
smartquestion 6 months ago prev next
I've been hunting for an effective method for real-time plant disease detection. I think I just found what I was looking for. Bravo!
johnsmith 6 months ago next
We're glad you find it effective for your needs. Feel free to reach out if you need any help or tips to get started.
technshader 6 months ago prev next
Which Raspberry Pi model did you use for this project? A Zero or 4?
johnsmith 6 months ago next
We used a Raspberry Pi 3 B+ for this project. The performance has been quite good so far.
codehiker 6 months ago prev next
How much time did you spend on training your TensorFlow model?
johnsmith 6 months ago next
We trained our model using a GPU farm, reducing the training time significantly. A quick estimate would be around 4-5 days for a decent model.
devmom 6 months ago prev next
My kid would LOVE this as a school project. Any plans to make a simplified version for educational purposes?
johnsmith 6 months ago next
That's a great idea! We've considered making a simplified version for educational purposes. Stay tuned - we'll announce it on Github.
curiousgeorg 6 months ago prev next
Do you have any literature recommendations for those interested in using TensorFlow for plant classification?
johnsmith 6 months ago next
Here are some good starting points: 1) TensorFlow research papers: <https://www.tensorflow.org/resources/paper-recommendations/2) TensorFlow tutorials: <https://www.tensorflow.org/tutorials>