It is a wearable device. It has a camera, and it detects all living and non living object. This module detects moving object also. It is made with raspberry pi 3, and a camera. One headphone connect with raspberry pi. When this module detects items, it gave a sound output through headphone. Hence the blind man know that item, which is in-front of him or her. We made it in very low budget, and it is very helpful for visually challenged people. And the Blind stick help him to detect obstacles.
Call Girls In Bangalore ☎ 7737669865 🥵 Book Your One night Stand
Development of wearable object detection system & blind stick for visually challenged people
1. DEVELOPMENT OF WEARABLE OBJECT
DETECTION SYSTEM & BLIND STICK
FOR VISUALLY CHALLENGED PEOPLE
Mentor: Prof. Chandrima Roy (Asst. professor , Dept. of Electronics and
Communication Engineering)
Heritage Institute of Technology, Kolkata
Team:
Moumita Majumder(1552210)
Arkadev Kundu(1552213)
Sanjib Basak(1552214)
Soumyadip Debnath(1552215)
2. PARTS OF THE PRESENTATION
Objective
Plan Of Action
Literature Study
Development Of Wearable Object Detection System & Associated Algorithm
Circuit Fundamentals
Basic operation
Circuit analysis
Results & Discussion
Development Of Blind Stick For Visually Challenged People
Circuit Fundamentals
Block diagram
Basic operation
Circuit analysis
Application
Challenges
Future Scope
Conclusion
Reference
3. OBJECTIVE
Design a module which can be helpful for visually challenged
people.
Develop an application for blind people to detect the objects
in various directions.
We are taking help of angle where object is placed to give
direction about the object.
Design a module which can give the sound output to identify
the object as well as human being.
4. PLAN OF ACTION
Objects Detection
Building Models Database
Objects Recognition
Literature Study
5. LITURATURE STUDY
1) Nowadays, the wearable monitoring system is the main
application of machine learning Likewise lots of wearable
devices are designed for visually impaired people. Few systems
are discussed here
2) Sensor assisted stick for the blind people describes about a
wearable equipment which consists of a light weight blind stick
and the obstacle detection circuit based on a sensor.
3) An innovative stick is designed for the visually disabled
people for their easy navigation.
6. LITURATURE STUDY
4) Multitasking stick is designed to indicate safe path to visually
disable people. The micro-controller based automated hardware
allows a blind person to detect obstacles in front of them. The
hardware part consists of a micro-controller which was incorporated
with an ultrasonic sensor.
7. DEVELOPMENT OF WERABLE
OBJECT DETECTION SYSTEM &
ASSOCIATED ALGORITHM
To understand Object Detection we have to understand some
techniques :
Bounding Box- The bounding box is a rectangle drawn on the
image which tightly fits the object in the image. A bounding box
exists for every instance of every object in the image. For the box,
4 numbers (center x, center y, width, height) are predicted.
8. Classification + Regression : The bounding box is predicted using
regression and the class within the bounding box is predicted using
classification.
9. Unified Method : This the most used method for object detection.
There are two architecture YOLO(you only look once) & SSD(single
shot detection). YOLO is the most advanced architecture but for
limited resources SSD architecture is most efficient
SSD architecture : Single Shot Detection architecture
10. Sl no.
1)
2)
Specification Image
1) CPU : ARM Cortex-A53 64 bit.
2) CPU Clock : 1.4GHz
3) No. of cores : 4 cores
4) RAM : 1GB
5) Memory : microSD
6) Ethernet : Gigabit over USB 2.0
(Max 300Mbps)
7) USB Port : 4
8) HDMI : 1
9) WIFI : 802.11 b/g/n/ac
10) Operating system : Raspbian
11) Power rating : 1.13A / 5V
Raspberry pi 3B+
1) Resolution : 5 MP
2) Max Frame Rate : 1080P 30fps
3) Lens size : 1/4”
4) Picture Format : JPEG,PNG
5) Weight : 5g
Pi-Camera
CIRCUIT FUNDAMENTALS
11. BASIC OPERATION
Step 1 - Pi camera Module takes a picture.
Step 2 - The image converts into edgy B&W picture.
12. BASIC OPERATION
Step 3 - In image processing, the features of the object
extracted(centre x, centre y, height, width), and checks it with
the previously given data set(Image processing done in
OpenCV).
Step 4 - If matches to any of the previously given data , then it
draw a bounding box around the object & and give the
confidence percentage.
Step 5 - Sound output.
13. CIRCUIT ANALYSIS
The circuit connection is very easy, as we shown below in the
image. The raspberry pi is connected to a mobile or laptop through
Portable hotspot connectivity, by that laptop or phone, it can be on
or off
15. Precision Chart :
Class Average Precision
Person 0.946
Bottle 0.659
Bird 0.712
Cat 0.936
Tv 0.856
Dog 0.891
Clock 0.719
Class Average Precision
Laptop 0.864
Vass 0.874
Suitcase 0.691
Book 0.628
Mouse 0.764
Table 0.735
Bicycle 0.714
16. DISCUSSION
1. Generally it is very important in Deep Learning to have training data
that is representative for your given problem.
2. Regarding image size and resolution: The smaller the image, the
faster will be your training but the accuracy will likely drop, because
your image will loose information.
3. It also depends how large the objects are that you want to detect
(few pixels or hundreds of pixels). Overlapping boxes should not
matter.
4. Some time when the object is not properly in camera frame, the
detector detects it wrong.
5. We have limited resource, we use raspberry pi 3B(1GB ram), for
heavy work like object detection, the hardware got heated.
21. BASIC OPERATION
Ultrasonic sensor sends high frequency pulses, these pulses reflects
from object and takes as Echo.
The speed of sound is 341 meter per second in the air, and the
distance between sensor and object is equal to time multiplied by
speed of sound divided by two.
Distance = (Time * Speed Of Sound) / 2
After the distance measurement, Arduino makes a beep format using
buzzer, when distance is high, frequency of beep is decreased and beep
frequency is increased when distance is low.
The range of HC-04 Ultrasonic sensor can measure 400cm in open
space, for more distance many other powerful sensors are available in
the market.
23. CIRCUIT ANALYSIS
The main component used for this device is the ultrasonic
sensor.
The ultrasonic sensor transmits a high frequency sound pulse
and then calculates the time to receive the signal of the sound
echo to reflect back.
The sensor is calibrated according to the speed of the sound in
air.
With this calibrated input, the time difference between the
transmission and reception of sound pulse is determined to
calculate the distance of the object.
24. APPLICATION
A well known application of object detection is face detection,
that is used in almost all the mobile cameras.
A more generalized (multi-class) application can be used in
autonomous driving where a variety of objects need to be
detected.
Also it has a important role to play in surveillance systems.
The main application of blind stick for blind people to detect the
obstacles in various directions, detecting pits and manholes on
the ground to make free to walk.
Object detection system can be used for tracking objects and
thus can be used in robotics and medical applications.
25. CHALLENGES
This device has been proposed with
the aim of developing a low cost
model.
Detection between object
background and object.
The main problem to detect the
object when the multiple objects are
present in a single frame due to
various dimension of object.
26. FUTURE SCOPE
This device can perform different task during the
development of the algorithm.
These device is more suitable if it is possible to get higher
frame per second of the system.
Further modifications to enhance the performance of the
system will be added. These include: A global positioning
method to find the position of the user using the GPS.
GSM modules to communicate the location to a relative or
care giver.
27. CONCLUSION
We successfully achieved the aim(object detection &blind stick)
we have proposed objects recognition approach for helping
blinds people.
This can be used in real-time applications which require object
detection for pre-processing in their pipeline.
An important scope would be to train the system on a video
sequence for usage in tracking applications.
28. [1] Ross Girshick, Je_ Donahue, Trevor Darrell, and Jitendra
Malik. Rich feature hierarchies for accurate object detection and
semantic segmentation. In The IEEE Conference on Computer
Vision and Pattern Recognition (CVPR), 2014.
[2] Ross Girshick. Fast R-CNN. In International Conference on
Computer Vision (ICCV), 2015.
[3] Shaoqing Ren, Kaiming He, Ross Girshick, and Jian Sun.
Faster R-CNN: Towards real- time object detection with region
proposal networks. In Advances in Neural Information
Processing Systems (NIPS), 2015.
[4] Joseph Redmon, Santosh Divvala, Ross Girshick, and Ali
Farhadi. You only look once: Unified, real-time object detection.
In The IEEE Conference on Computer Vision and Pattern
Recognition (CVPR), 2016.
REFFERENCE
29. We try to import some ray of hope to their life
THANK YOU