cash inbox smart cash assistant for blind6894

Project TitleProject Area of Specialization Artificial IntelligenceProject Summary

Our Final Year Project, which will be a product named “CASH-INBOX” will be a smart device that will assist blind people with cash recognition problem in their daily lives. Blind people commonly become a victim of cheating related to cash bills because they are visually impaired. Accurate cash identification for the blind is a serious issue because such monetary loss can directly affect their economical status and living. Any kind of fraudulent activity regarding cash with the blind person can affect his/her living. In Pakistan, anyone can become a victim of such fraud. Any business owner or a representative of a cash paying body can execute such fraud against anyone. Blind people are the easiest target for individuals who execute such fraudulent activities because of their inability of to recognize cash.

To address this problem, our project solution is to implement a smart hardware device to assist a blind person with cash recognition to prevent such fraud. This device will recognize each Pakistani currency cash bill and provide audible feedback about the cash inserted, to the blind person via hands free, wireless ear buds or speaker. This will help the blind person to easily acknowledge/evaluate the cash amount returned to him by a shopkeeper for example. We will be using Raspberry Pi along with a camera module to obtain the picture of cash note inserted and incorporate Machine Learning techniques for cash bill recognition.

According to statistics of 2013, well above two million people suffer from blindness in Pakistan. Initially, these will be our targeted potential users of our project. Considering future scope of our project, we will scale and deploy our project solution internationally by adding support for other currencies because according to World Health Organization, more than 39 million people in the world are suffering from blindness.

Project Objectives

The main purpose and motivation of implementing this project is to prevent cash bill fraud against blind people. The importance of implementing this project for the blind is that they commonly become a victim of cash bill fraud due to their visual impairment and inability to recognize cash by other means. This project will be smart cash assistant for the blind. This project will be in the form of a portable device that the blind people can incorporate it in their daily lifestyle and make use of it for Pakistani cash bill recognition while on the go. The final product of our project solution will be a prototype having the following characteristics:

1. Pakistani cash bill recognition.
2. Audio output stating the worth/amount of each cash bill in the form of voice to the blind user.
3. Audio output stating the total amount of all cash bills recognized in a particular session in the form of voice to the blind user.
4. Audio output stating if the inserted cash bill is invalid in the form of voice to the blind user.

Project Implementation Method

Our project is implemented using both hardware and software technologies. The final project prototype will be hardware and software integrated system.

Software implementation methodology comprises of two major modules: Machine Learning model and Audio voice generator. The machine learning model will be trained using the Pakistani currency data set containing the images of cash bills. The picture of inserted cash bill is preprocessed using Digital Image processing techniques. Then, this preprocessed image is passed to the code containing the trained machine learning model for evaluation. This machine learning model will recognize the cash bill and classify its worth/value. Then, this result will be passed to the Audio Generation code module which will generate an audio message containing the result in English (Urdu TBD) voice form. All the software will run on raspberry pi 4B which will be running a variant of Linux operating system.

Hardware implementation methodology makes use of Raspberry Pi 4B. When cash bill is inserted, IR sensor is used which checks the presence of cash bill and changes the sleep mode of the device to active mode. The Pi camera module is used to take picture of this bill and passes its image to the software module which is being hosted on the raspberry pi. The audio result is generated by the software module and it is sent to the ears of blind user via hands-free connected with the Pi audio jack, wireless ear-buds connected to Pi via Bluetooth or speaker. The cash bill will be ejected by the CASH-INBOX device by using belts, pulleys and servo. This was for each individual cash bill inserted. When all cash bills have been recognized, the blind person will tap a Push Button located onto the device chassis, this Push Button will be designated to terminate the recognition session. The device will provide the total amount of all the cash bills inserted via audio and then the device will go into sleep mode to save battery power. All these hardware components will be connected with raspberry pi using jumper wires and other types of wires if required.

Benefits of the Project

The main and foremost benefit of this project is that it assists the blind people with Pakistani cash bill value/worth recognition. Benefit of portability and recognition of cash bills on the go is also included. Another benefit to blind people by having this device is that they do not have to learn of practice other methods of identifying the worth/value of cash bills which are often hard to learn and fail at certain times in the real world. This makes the blind people to manage their cash assets independently.

Apart from blind people, this device can also benefit people suffering with other types of visual impairment to recognize cash. This device will also benefit people who suffer from multiple brain diseases to recognize cash bills.

The recognized cash bill value or worth is notified to the user via audio.

Technical Details of Final Deliverable

The final deliverable of our project will be a product prototype that is portable and can be used on the go. This will be a smart device that will use Artificial Intelligence and Internet of Things technologies. This will be a device that will recognize cash bills and provide audio feedback to its user. This device will enclosed in an external chassis and this chassis will contain buttons, cash insertion area and other ports for connectivity. The push buttons will be used for volume control and session termination. Ports will be used for connectivity. Audio jack port for connecting hands-free. The technical details of each component used in the device and its purpose are stated below.

The Final project prototype will use the following hardware components:

1. Raspberry Pi 4 B: Core hardware unit for computation, interfacing and hosting software.
2. Pi Camera module V2: For taking picture of cash bill inserted into the device.
3. Power Bank: For powering the device and enabling portability.
4. IR Sensor: For switching the device from sleep state to active state and vice versa.
5. LEDs: For illuminating the cash bill cabin for camera and power indicator.
6. Jumper Wires: For connecting all the internal hardware components together.
7. Push Button Switch: For increasing and decreasing audio volume.
8. ON/OFF Button rocker: For powering on/off the device.
9. Wireless ear buds/hands free/speaker: For providing audio based result to blind user.

The Final project prototype will use the following software components:

1. Python: As core programming language for development.
2. Data Set: containing all the pictures of Pakistani currency cash bills.
3. Linux: Base Operating System executing all the software and making it interfaces with hardware.
4. Open CV: Used for preprocessing the captured image of cash bill.
5. Machine Learning Model: Trained model that will classify and predict the cash bill.
6. Voice Feedback API: For generating audio based output in the form of voice.

Final Deliverable of the Project HW/SW integrated systemCore Industry ManufacturingOther IndustriesCore Technology Artificial Intelligence(AI)Other Technologies Internet of Things (IoT)Sustainable Development Goals Industry, Innovation and Infrastructure, Reduced InequalityRequired Resources

Item Name	Type	No. of Units	Per Unit Cost (in Rs)	Total (in Rs)
			Total in (Rs)	36343
Raspberry Pi 4-B (2GB RAM)	Equipment	1	11000	11000
8 MP Pi Camera module V2	Equipment	1	5250	5250
Case for Raspberry Pi 4	Equipment	1	700	700
Mi Power Bank 3 (10000 mAh)	Equipment	1	4000	4000
IR Transmitter and Receiver Diode	Equipment	5	50	250
LED Strip 3V DC	Equipment	1	438	438
Jumper Wires (pack of 40 pcs)	Equipment	1	240	240
Push Button switch (pack of 5 pcs)	Equipment	1	65	65
ON OFF Button rocker switch (pack of 5 pcs)	Equipment	1	100	100
Wireless Earbuds	Equipment	1	2000	2000
Servo motor	Equipment	2	650	1300
Device Chassis	Equipment	1	10000	10000
belts and chain etc	Miscellaneous	1	1000	1000