Fault Detection in Monochromatic Garments using Image Processing

Project Title

Fault Detection in Monochromatic Garments using Image Processing

Project Area of Specialization Electrical/Electronic EngineeringProject Summary

This project is about fault detection in monochromatic garments using image processing. In industries fault detection is being carried out on raw fabrics. Also, inspection is done manually. But in the mentioned project faults are being detected in garments rather than raw fabrics. Here detection is carried out automatically instead of manual inspection. Moreover, the key motivation is that this project detects fault that is not visible to human eye. In addition to this the system will be portable. Ky purpose of doing this project is to inspect the garments automaticallyratherthan manually which leads to the enhancement of the quality of the product, hence decreasing the error and time required for manual inspection. It is carried out in such a way that garment is placed on dummy and is being hung on the rod of the wooden structure. Four cameras being mounted on the structure acquire the images from different angles in the presence of proper lights being adjusted on the structure. Than image processing techniques are being applied to the acquired images nd in the end signal is being sent to the robotic arm. The robotic arm than takes the decision for the placement of the garment in the respective boxes depending upon the signal being recieved.

Project Objectives

The main objective of this project is to find four kinds of faults in monochromatic garments including stain, color fadding, hole and ladder (not visible to human eye). Moreover, system will be made in such a way that it will be portable. Key purpose of doing this project is to inspect the garments automatically rather than manually which leads to the enhancement of the quality of the productand hence decreasing the error and time required for manual inspection.

Project Implementation Method

For the formation of a system that will be used for detection of faults in monochromatic grments, first, monochromatic garment will be given as input by hanging the garment on the specially designed dummy and will be placed on the rod that will be mounted on the woiden support structure. Afterwards its image will be taken by four Logitech cameras mounted on the structure. For getting better results of image acquisition, proper lightning system will be designed and will be adjusted on the body of the structure. Furthermore, Raspberry Pi will convert the colored image of the garment into gray scale image. After the conversion its histogram will be formed and it will be equalized as well. In addition, feature extraction will be done. In the end on the basis of processed image robotic arm will take the decision and will place the garment accordingly in the respective box placed for four kinds of faults including hole, stain, ladder and color fading. There will be a total 6 boxes, one for each fault, one for garment having no faults and one for garment having multiple faults.

Benefits of the Project

This project helps in autonomous quality assurance of the garments. Moreover human error is removed. In addition to this seperation will be carried out on the basis of error. Also, it helps in reduction of time consumption taken for detection of faults. As it helps in the enhancement of the quality of the garments, this will result in more profit.

Technical Details of Final Deliverable

Firstly garment will be placed on a specially designed dummy (made of light plastic). It will than be hanged on the rod of support structure. Support structure is being made of wood weighing almost 35 kg. Moreover, image accquisition will be done using four Logitech cameras and proper lighting will be used for better results of image accquisition. All image processing will be carried out using Raspberry Pi. After which it will send signal to robotic arm. Robotic arm will be designed according to the constraints of this project. Robotic arm will than place it in its respective box. 

Final Deliverable of the Project HW/SW integrated systemCore Industry OthersOther IndustriesCore Technology Wearables and ImplantablesOther TechnologiesSustainable Development Goals Decent Work and Economic Growth, Industry, Innovation and Infrastructure, Responsible Consumption and ProductionRequired Resources

Item Name	Type	No. of Units	Per Unit Cost (in Rs)	Total (in Rs)
			Total in (Rs)	70600
Support Structure	Equipment	1	8800	8800
Raspberry Pi	Equipment	1	6000	6000
Memory Card	Equipment	1	1000	1000
Cameras	Equipment	4	3700	14800
Dummy	Equipment	1	6000	6000
Lights	Equipment	2	8000	16000
Robotic Arm	Equipment	1	15000	15000
Reports	Miscellaneous	5	600	3000