Overview / Usage

This project uses YOLO object detection to identify and remove weeds in agriculture. It offers an efficient, sustainable alternative to traditional pesticide-based weed management, resulting in increased yields. This system can automate weed detection and remediation, making it useful for farmers.

Methodology / Approach

Our project focuses on using technology to solve the problem of weed Classification in agriculture. To achieve this, we used the YOLOv3 object detection model and implemented it using the TensorFlow platform. To enhance the training efficiency of our model by trimming down the original architecture to achieve better accuracy.

We followed standard techniques and best practices in the development process, such as data pre-processing, model training, and evaluation. We created customized YOLOv3 from scratch as per our requirement.

In summary, our methodology involved leveraging state-of-the-art deep learning technique, Intel OpenVINO toolkit and TensorFlow framework to develop an efficient and accurate weed detection model that can be deployed in real-world agricultural settings. By using advanced technology, we aim to offer a sustainable and cost-effective solution to weed Detection, ultimately helping farmers improve crop yields and profitability.

Technologies Used

Technologies:

Deep Learning

YoloV3

Library:

Tensorflow

openCV

OS:

Linux

Windows

Hardware:

CPU Configuration: RAM: 32 GB

                                CPU :  11th Gen Intel(R) Core(TM) i7-1165G7 @ 2.80GHz   1.69 GHz 

                                HDD : 500 GB

GPU Configuration: RAM : 32 GB

                               GPU : NVIDIA  GeForce RTX 4090 - 16 GB (2 GPUs)

                                HDD : 500GB

​

Intel technologies

OpenVINO

MKL

GAL

​

Repository

https://github.com/payal211/weed-detection