Safai Dost

Project Title

Safai Dost

Project Area of Specialization Artificial IntelligenceProject Summary

A system which can be easily help every citizen to clean their city using deep learning concepts ,this project is basically an app which allows user to to sign up with the credentials like name age cnic number etc. for authentication then user is redirected to complain portal clicks the picture of the trash and upload on the server the server predicts whether the picture is trash or non-trash ,then after the success the complaint is registered with location pin point which helps to form the graph in the map ,also it helps the trash picker’s to pick according to the points with database respectively.Trash is the common problem in the world especially in the under developed countries like Pakistan.In Karachi city having five districts creates the hard situation to manage the things ,we see our streets dirty ,unclean which causes problems. Having this problem requires a good idea for revolutionary process, so using deep learning and some other skills we can solve this main problem including other places then this city.

Project Objectives

The purpose of this project is to reduce human efforts and to ensure the mangement of the organization by saving time and money.

Our main focus of our project is on the following objectives:

• The proposed framework would have the option to computerize the strong waste checking procedure and the management of the overall assortment process using log based system.

•  In the proposed framework, at whatever point the waste receptacle gets filled this is recognized by the program, which transmits it through pin point to the collector at the ideal spot in the region or spot.

• In the proposed framework, the got signal shows the waste area status at the observing and controlling framework.

• The program is also capable to help the truck drivers in order to garbage collection by making their routes more shorter and time saving.

Project Implementation Method

Using CNN with proper dataset more than 1000+ images per class we train the model using sequential method that helps to predict the status of the image either it's trash or non-trash ,than this model file with extension .h5 used in the android application with firebase ,the picture from android captured hit the model file predicts the status of it,then all the data is stored in the database with clicked picture,person cresdentials and  with a pin point  generate.Using Google Map API a proper sets of points is set up on the map just like we saw now a days in covid-19 issue.

Now after a points implemented every driver has its own area to clear according to the points,an algorithm here is used which is travel salesman problem allows to save time and money(fuel) .

Benefits of the Project

For identification of the trash procedure the system will be utilized in several areas which is “Trash Recognition” system. Some of its uses are :

• As a Good Citizen (Rank Based System) 
• Covering the Bad Sights
• Cleaning Process
• Tracking the Location of  Affected
• Maintaining Logs 
• Fuel and Time Save
• Data Handling
• Trash Count

Technical Details of Final Deliverable

Basically its a project based on two android applications:

1. For Users
2. For Trash Truck

Users: have an access to upload the affected area picture (after sign- in) then system will tell them about the status if it is not trash than user have to submit the right picture else it will save the image in database and marked a pin point of the area in the map (along with altitudes and latitudes) and save it in logs directory with person cresdentials,the more the accurate reports in person log the more towards the rank system(Good Citizen) .

Trash Truck : have an access to visit specific blocks/areas which is fixed by the organization and after the approval have to visit the affected area for cleanliness ,driver has to upload the picture of  cleared area thats just saves in the log file ,the less number of complains leads to more power to the driver reputation.

Final Deliverable of the Project Software SystemCore Industry OthersOther Industries Petroleum , Transportation , Health Core Technology Artificial Intelligence(AI)Other Technologies OthersSustainable Development Goals Good Health and Well-Being for People, Decent Work and Economic Growth, Life on LandRequired Resources

Item Name	Type	No. of Units	Per Unit Cost (in Rs)	Total (in Rs)
			Total in (Rs)	46084
GPU	Equipment	1	14000	14000
API	Equipment	1	32084	32084