Good roads and feasible traffic conditions are both essential for a stress-free driving experience. Similar to many other developing nations; driving conditions are not feasible in various cities of Pakistan, due to the existence of potholes, sharp drop-offs and boundaries that are not sign-posted.
ROADVISION
Good roads and feasible traffic conditions are both essential for a stress-free driving experience. Similar to many other developing nations; driving conditions are not feasible in various cities of Pakistan, due to the existence of potholes, sharp drop-offs and boundaries that are not sign-posted. Drivers are required to practice outrageous alert. This situation further deteriorates when the monsoon season hits the country because in majority areas water drainage is inadequate as a result water gets accumulated on the roads. Quite frequently these broken drains and uncovered potholes become the cause of accidents. Other than this, traffic congestion is a problem worldwide, although there exist many driver assistance systems but they are either out of reach of the general public or less efficient.
In this project we present a system named Road vision that can be used for analyzing the current vicinity and alert the driver. It includes a raspberry pi module accompanied with a web cam; for detection of road, GSM Device; for sending notification to other users having the same product, and Pi screen; is connected to camera, it is basically used for displaying the road condition to the user when any pothole, conjunction or lane is detected. Concurrantlly screen will display alternate routes in case of any damaged road and congestion ahead of the driver’s opted route.
this project would assist drivers:
1. by detecting potholes on road.
2. by defining the cause of congestion .
3. by generating the cause of congestion to other cars having the
same systems.
4. by detecting the traffic lights.
5. by generating an alert via pi screen and audio alert .
6. while doing all these task ,updating other driver by using GSM
device.
7. Create a data repository for future analysis
we are providing a user with display screen and audio alert generator to avoid collision, potholes, and avoid traffic congestion situation at the same time the alert will be generated in all cars having road vision, through map. for that we are using hardware like
Hardware Side:
1) Raspberry Pi:
The execution of our item is on the dashboard of car,where we have restricted space ,so we can not utilize equipment like PCs there,but we can not disregard the consideration of figuring power require for picture handling ,so to fulfill the both situation we choose to utilize edge PC like raspberry pi.
2) Camera:
The fundamental reason for this task is to actualize AI on machineutilizing camera as sensor to see the environment.
3) GSM Device:
GSM stand for (global system for mobile communication)..We are utilizing GSM gadget for sending warning to other client having the equivalent product.
4)Pi Screenpi:
screen is associated with camera utilized in this undertaking .it isfundamentally utilized for showing the street condition to the clientwhen any pothole , combination or path is identified, simultaneously the alarm is created in different gadgets through which client can seethe reason for conjunction.
Software Side:
Equipment endorsed above is nothing without programming ,all the usefulness equipment performs depends on programming side of our module.
linux bolsters ”python programming language” which is the most dominant language for man-made brainpower related operations,which additionally incorporate ”picture preparing”.
Image Processing:
python supports various power full libraries like keras,opencv andtensorflow,which plays a vital role in image processing tasks. like;
Opencv:
OpenCV is a library of programming capacities or function mostly focused on constant PC vision.
In this project opencv helps us on performing camera operations,like capturing of frames and performing pre-processing tasks.
keras:
it itself provides number of pre-trained models trained on differentmachine learning algorithms.
keras pre-trained models can be used as it is or trained on ourown objects by ”transfer learning ”approach,.
that’s why we use keras to import our preferred model ”mobile-net” and apply transfer learning on it.
Tensor Flow:
Tensor flow also provides models with their configuration files,whichhelps in training the model on our own data sets.
We use this library to train ”mobile-net SSD” on our own dataset.we particularly use this approach on training of ”potholes”
this project would help car drivers in better road travel towards there destination.it would alert them about what coming ahead of them
,wether there is any traffic congestion ? ,is road damaged? or is there any expected collision?
this project would help in minimize the road hectics,so that driver can dolow the beter routess twards their destination, with better perception of road enciroment.
| Item Name | Type | No. of Units | Per Unit Cost (in Rs) | Total (in Rs) |
|---|---|---|---|---|
| raspberry pi 3b | Equipment | 1 | 10987 | 10987 |
| pi camera | Equipment | 1 | 3700 | 3700 |
| gsm module | Equipment | 1 | 2500 | 2500 |
| memory card | Equipment | 1 | 640 | 640 |
| screen for raspberry pi | Equipment | 1 | 4500 | 4500 |
| neural compute stick | Equipment | 1 | 23000 | 23000 |
| Total in (Rs) | 45327 |
The existing systems consist of an electronic energy meter/electro-mechanical meter fixed...
MobiTest is a codeless automation tool specifically for mobile applications. It can be use...
The Bahadur bot function is to help people who are stuck or need of help. Human dete...
Our application named ?Necto? would open up the earning facilities for the Chefs & del...
The Global Positioning System (GPS) is comprised of a network of satellites that continual...