You are on page 1of 12

<Industrial IoT>

<DEFENDERS>
<BANNARI AMMAN INSTITUTE OF TECHNOLOGY>
<DEFENDERS>

• BARATH P
• B.E II YEAR
• IVth SEMESTER
• COMPUTER SCIENCE AND ENGINEERING
• PARTICIPATED IN VISAI(Vel Tech Hackathon), Digital Gov Hack (Saudi Arabia), SRIT
HACKATHON(Coimbatore) and TNSDC Naan Mudhalvan Hackathon
• PROJECTS: VIRTUAL LIBRARIAN, AI-BASED SMART SURVEILLANCE SYSTEM, MOVIE
RECOMMENDATION SYSTEM, VOICE CONTROL-BASED COMMAND INSTRUCTION SYSTEM
<Industrial Iot>

• No Of Challenges undertaken: All


• Broad description of the proposed solution: Voice-based instruction mechanism
• It can be achieved through Natural language Processing and Voice Recognition
• The First step is to collect the data from various sources and Preprocess the data
• The next step is to develop an NLP model that can understand natural language
commands and queries. BERT is a transformer-based model used for natural language
understanding and question-answering. It understands user queries and commands and
generates relevant visualizations based on the input data.
• Once the NLP model is developed, it can be integrated with Power BI or similar tools using
APIs or other integration methods. This will allow the system to generate graphs and
charts based on user queries.
<Industrial Iot>

• No Of Challenges undertaken: All


• Broad description of the proposed solution: Cracks Identification using Computer Vision
• It can be achieved through Computer Vision
• The initial step is capturing images of large objects from different angles and annotating
the images with labels indicating the location of any cracks or dents. Collect and
Preprocess the data
• Next step is to develop an image processing pipeline that can extract features from the
images and identify any cracks or dents. This can be done using a combination of feature
extraction, and edge detection.
• It can be integrated with a back-end server that can store the images and highlight the
areas of any identified cracks or dents.
<Industrial Iot>

• No Of Challenges undertaken: All


• Broad description of the proposed solution: Dimension measurement and quality
inspection through the mobile camera.
• Computer vision can be used to measure dimensions and analyze images captured by the
camera.
• The system can use object detection to detect simple objects such as squares, triangles,
and polygons in the image.
• Once the object is detected, the system can use computer vision techniques such as
Scale-invariant feature transform (SIFT) to calculate the dimensions of the object
accurately. This can be done by using the scale set by the user, such as mm or cm, to
convert the pixel measurements of the object into real-world dimensions.
• The system can also perform quality inspection on the object by analyzing its shape, color,
texture, or any other visual characteristics.
<Challenge 1>
• Voice-based instruction mechanism
• To develop an NLP model that can understand natural language commands and queries.
BERT is a transformer-based model used for natural language understanding and
question-answering. It understands user queries and commands and generates relevant
visualizations based on the input data.
• Once the NLP model is developed, it can be integrated with Power BI or similar tools using
APIs or other integration methods. This will allow the system to generate graphs and
charts based on user queries.
• A laptop or a desktop computer with a microphone and speakers can be used for this
purpose.
• Using Google Speech Recognition API (Google Cloud Speech-to-Text API) it interprets
voice commands and generates visualizations in PowerBI.
Block Diagram

NLP (BERT)
Data Retrieval and
Voice Recognition Natural language
analysis
Processing

PowerBI Integration
<Challenge 2>

• Cracks Identification using Computer vision


• The first step is to collect and preprocess the data that will be used to train the model.
This can be done by capturing images of large objects from different angles and
annotating the images with labels indicating the location of any cracks or dents.
• Next step is to develop an image processing pipeline that can extract features from the
images and identify any cracks or dents. This can be done using a combination of feature
extraction, and edge detection.
• Edge detection is a common image processing technique used in computer vision that
detects and highlights the boundaries of objects in an image.
• Feature extraction algorithms are used to extract relevant information or features from an
image that can be used to identify cracks or dents.
• Once the model is trained, it can be integrated with a back-end server that can store the
images and highlight the areas of any identified cracks or dents. This can be achieved
using APIs or other integration methods.
Block Diagram

Image
Image capture Feature Detection
Preprocessing

Integrate with Back-


Store the images End server through Edge detection
API
<Challenge 3>
• Dimension measurement and quality inspection through the mobile camera.
• Computer vision can be used to measure dimensions and analyze images captured by the
camera.
• Before capturing images or videos, the camera needs to be calibrated to ensure that the
measurements taken are accurate.
• Once the camera is calibrated, the system can use object detection to detect simple
objects such as squares, triangles, and polygons in the image.
• Once the object is detected, the system can use computer vision techniques to calculate
the dimensions of the object accurately.
• SIFT is a feature extraction technique that can be used to identify key points in an image
and describe their local features. By matching these features across multiple images, the
system can estimate the size and orientation of an object and calculate its dimensions.
• The system can also perform quality inspection on the object by analyzing its shape, color,
texture, or any other visual characteristics.
Block Diagram

Image
Mobile Camera Image Acquisition Preprocessing and
Feature extraction

Dimension
Result Display Quality Inspection measurement(SIFT)
Mistral Solutions Pvt. Ltd. AXISCADES Technologies Limited
No.60, 'Adarsh Regent', 100 Ft. Ring Road, Domlur Ext, Kirloskar Business Park, Block ‘C’, 2nd Floor, Hebbal,
Bengaluru - 56 0 071, India Bengaluru – 560 024, India

Tel: +91-80-4562 1100 Tel: +91 80 4193 9000 | Fax: +91 80 4193 9099
www.mistralsolutions.com | info@mistralsolutions.com www.axiscades.com | info@axiscades.in

You might also like