Skip to content

Trained a yolov5 with available dataset in roboflow thanks for the public dataset

Notifications You must be signed in to change notification settings

jayaprakashll/transmission-line-inspection-yolov5

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

8 Commits
 
 
 
 

Repository files navigation

the output of the model the Map of the model is 92.87%

Insulator Defect Detection System for Power Transmission Lines

This project introduces an innovative approach for detecting insulator defects on power transmission lines using advanced AI technology.

Overview

The system utilizes the Jetson Nano platform mounted onto a drone for efficient aerial inspection of transmission lines. This setup enables high-resolution imaging and real-time defect identification.

Features

1.Drone-based Inspection: Drones equipped with Jetson Nano conduct aerial inspections of transmission lines.

2.AI-powered Defect Detection: Sophisticated algorithms accurately identify and classify insulator defects.

3.Geofencing Technology: Precisely delineates inspection areas for comprehensive coverage.

4.Web Portal Integration: Detected defects are automatically pinpointed and transmitted to a web portal for analysis.

5.Efficiency and Risk Mitigation: Enhances inspection efficiency and ensures the integrity of power transmission lines by promptly detecting defects.

Implementation

The system combines AI technology, edge computing, drone-based inspection, and geofencing to facilitate proactive maintenance of power infrastructure.

for training the dataset has been taken form

we trainned the model in the google colab and the link is provided

Steps to run locally

step1: clone the repostory:

step2:open the cloned repo in vscode

step 3: first install the requirement files run the code on the terminal of the vs code : pip install -r requirements.txt

step 4: the following commands are used to run the model

to acess the web cam : python detect.py --weight best.pt --source 0

--weight = the trained weight file that we are using --detect python code that used to run the weight file --source the resources that we are goign to use

sorce 0 indicatest the web cam acess after opening the web cam place the insulators photos near the web cam , to see the identify the defects

step 5: to access the folders that contains of the images that

run this code code : python detect.py --weight best.pt --source testing 'here testing is a folder name that contains the images of the insulaotres

step 6 : we can aslo use viedos for the model

About

Trained a yolov5 with available dataset in roboflow thanks for the public dataset

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published