Geometric Parameters Measurement of Optical Fiber by Image Processing

The aim of our project mainly focus on quality inspection of a product by measuring its dimensional parameters, which in our case its optical fiber. To do quality inspection we will be creating our machine vision system in such a way that it can inspect the cladding diameter and number of cores of o

Project Title

Geometric Parameters Measurement of Optical Fiber by Image Processing

Project Area of Specialization

Mechatronics Engineering

Project Summary

The aim of our project mainly focus on quality inspection of a product by measuring its dimensional parameters, which in our case its optical fiber. To do quality inspection we will be creating our machine vision system in such a way that it can inspect the cladding diameter and number of cores of optical fiber as required. We will be utilizing a lighting controlled environment where our camera can take a quality picture of optical fiber with negligible noise. We will utilize the illumination techniques in our arsenal to create that environment for our machine vision system. After creating the environment we will be using the OpenCV library to measure the parameters of our optical fiber. The image processing techniques is applied to get the required parameters.

Project Objectives

• To design a measurement system for geometric measurement of optical fiber

• To fabricate the prototype system as per design

• To develop an algorithm for geometric parameters measurements by image processing

• To do Experimental testing of the prototype system

Project Implementation Method

This project core objective is quality inspection of fiber in which our focused area are cladding and number of cores of the optical fiber.

First, we developed the project algorithm using OpenCV library in python to complete the assigned task. The image processing techniques that we will be using are: reading an image, converting image to grayscale, applying Gaussian blur method to remove noise from the image, applying contour method to get the internal and external contours. We draw contours after detecting the circle which represents our number of cores and cladding of our optical fiber and further process it to get the required parameters. We also take a reference object whose dimensions will be known. In our code we will detect the reference object and utilize it to measure our cladding diameter and observe if the result is as required.

Then, we designed our prototype in Solidworks, in the design the fiber is fixed at a certain location and an adjustable camera is placed in front to capture its image. Front lighting and Back lighting is used to focus on our object and remove shadows respectively.

Next, the design is fabricated with wooden material.

Finally, we perform some tests and get several outputs of which we observed the best way of implementation.

Benefits of the Project

This project can be used in field of Quality Control and with little amendment we can measure several other parameters of any similar product such as length, width, diameter, height, roundness, etc.

This is an industrial project that is cost efficient as it will reduce manual labor and because machine vision systems are faster than humans it can check parts at a faster rate. The accuracy of the process helps to reduce the dispatch of defective parts to customers. This in turn reduces external failure costs such as replacements, loss of sale, and warranties.

It has several other benefits:

• Accuracy: Machine vision systems provide greater precision and accuracy of quality inspection when compared to human operators. Even when humans rely on a magnifying glass or microscope, machines are still more accurate because they can “see” and measure parts to a higher tolerance and the part in our case is fiber
• Speed: Optical fiber can be inspected at higher speed compared to inspection by human operators. In addition, such high-speed inspections can be carried out with greater efficiency and productivity.
• Repeatability: Quality inspections can be repeated in the exact same way by machine vision systems without fatigue, 24/7 and 365 days in the year. In contrast, human inspectors may arrive at different measures at different times, even if all the parts are exactly the same
• 100% Inspection: Machine vision systems can be designed to inspect 100% of the parts or products manufactured in our case its optical fiber. Otherwise it might only be possible to inspect parts/products that are randomly sampled using statistical methods and tools.

Technical Details of Final Deliverable

• In our machine vision system through trial and error, we use the illumination technique and developed a controlled environment in which we use backlighting to remove the shadows and front lighting to focus on our object.
• Based on our finalized illumination technique we designed a prototype to set up all the required components. Hardware is a wooden box with 2 attached led light and includes a sliding mechanism to adjust distance between the camera and fiber to get desired image.
• CMOS technology camera is used to take high quality image.
• In Jupiter Notebook, Python language is used with OpenCV library to measure the parameters of optical fiber.
• To get our final result, first we read the image taken by the camera, then convert it into grayscale, next apply Gaussian blur method to remove noise, after that we apply contour method to calculate the number of cores of the fiber, then using the reference object with known parameters we can easily calculate the parameters of the fiber.

Final Deliverable of the Project

HW/SW integrated system

Core Industry

Manufacturing

Other Industries

Education , IT , Medical , Media , Security , Telecommunication

Core Technology

Artificial Intelligence(AI)

Other Technologies

Others

Sustainable Development Goals

Quality Education, Industry, Innovation and Infrastructure

Required Resources

If you need this project, please contact me on contact@adikhanofficial.com