Happy to have comments and share.........................😄😄😄😄😄😄😄😄

CRACK DETECTION IN PIPE LINE

 

This blog comes for the people especially mechanical engineering students for the motive for development of Live tracking of crack as well as welding in a single robotic manipulator ( Live tracking using android devices )
Since i am only focused on making a prototype and due to lack of  financial support i couldn't  implement welding so i will give my full support to people who are interested to implement the idea 
A brief introduction about the project before starting
It has been a difficult task for the human beings to reach and weld the cracks and irregularities inside long tunnels and pipes. Welding outside the pipe is a widely accepted technique. Radiography is effectively employed for this purpose. But the crack need not extend to the outer periphery always. Thus the process would be more efficient if it can weld cracks inside the pipes also. It is difficult to do it manually. A robot designed to detect and weld the cracks can solve this problem. A robot can be used for inspection of cracks in a tunnel or pipe in which maintenance cannot be done manually. If the pipe carries suffocating gases or is smaller in dimensions, then a similar difficulty arises. This project deals with the design and fabrication of a vision based welding robot to detect and weld cracks inside such tunnels.

The operation mainly consists of:
 1) Sensing and processing the image.(Camera attached to the robotic manipulator )
 2) Image analysis ( Especially MATLAB R15 b)
 3)Processed image is send to android devices for further verification
 4) Welding Suitable gas welding can be preferred Refer Link( http://www.ehow.com/about_5474325_types-welding-gas.html).    

Project detailed view
A camera attached to the manipulator act as the vision system. It captures images continuously and sends it to the computer. MATLAB R15 b is used for image processing.

 Figure (1.0)                                              Figure (1.1)
figure 1.0  shows crack and the other (figure 1.1) without crack   
 Crack is detected in the digital image. After analysis different poses of points on the crack is obtained .Each image is processed using MATLAB and is checked for the presence of crack in the image taken by accessing camera attached to the manipulator (Note: Camera should be of good pixel for the clarity for efficient filtering of crack   )  .

For aligning arms in the proper direction signal is sent to the controller for controlling the servo motors by help of arduino board which can be easily accessed using MATLAB R15 b cost around 5 dollar
link to buy  arduino board in amazon market
(http://www.amazon.in/Arduino-ATmega328P-ATMEGA16U2-Compatible-Cable/dp/B015C7SC5U?tag=googinhydr18418-21&tag=googinkenshoo-21&ascsubtag=05a43871-1a57-4508-9e4c-58a79d41cc84)
 By using A stepper motor  DC electric motor that divides a full rotation into a number of equal steps  required orientation welding can be easily done  

Important Notes :

• FOR IMAGE ANALYSIS   

 Every image sent by the camera is processed in MATLAB and is checked continuously for the presence of crack. Once the crack is detected it is directly send to cloud for sending to the android device with help of MATLAB .
To know more about linking android device to MATLAB  you can refer this link 
 For Linking using Wifi You  can refer
 https://in.mathworks.com/help/supportpkg/mobilesensor/ug/set-up-and-connect-to-android-device.html?requestedDomain=www.mathworks.com
 For Linking using  cloud  You  can refer
 http://stackoverflow.com/questions/28358291/connect-android-application-with-a-matlab-application-on-server
Formula which help to create a algorithm  (Don't gets shocked you can get support from or you can take help from other computer science people from your college
 In our context a crack can be defined as the one which is darker and conspicuous than the background. Thus when a RGB image is converted to a Black & White image, portions of crack will be represented by black pixels. But the presence of a small number of black pixels cannot ensure the presence of a crack, because small group of isolated black pixels can be found due to the presence of noises and other disturbances in the image.
The crack detection algorithm is based on the sudden increase in the number of black pixels in an image. Basically, any image will be compared with an image without any crack based on the number of black pixels it has. It is thus essential to find the average number of black pixels that can be found in an image which is devoid of any cracks. To avoid the accumulation of unwanted black pixels, proper light settings are also arranged on the robotic manipulator. The lightnings provided ensure the presence of a constant intensity light to fall upon the surface of the tunnel while the image is captured. This helps in reducing the interference of the outside lighting conditions with image processing. After analyzing a number of images without crack, we can safely assume a threshold value for the number of black pixels, above which the image can be said to be containing cracks. With slight chances of errors we can approximately find the presence of crack inside the tunnel using this crack detection algorithm. This image will be finally analyses to obtain the positions of the crack.


 Best of Luck.....................

HAVE  A WILL THAT'S ALL YOU NEED TO COMPLETE YOUR PROJECT .....