Design and sorting of an object identification on machine vision by using line scan camera

This research represents designing a system to identify objects on a conveyor belt using machine vision. In the present study, a machine vision based on one line scan sorting was developed, the purpose being to sort objects based on various stages of maturity. Many different methods are available for object identification. But it design a system that separates and counting them. Different objects placed on the conveyor belt moves along, and a camera placed above the belt takes real-time video and feeds it to the MATLAB software for processing the object to compare with the basic template object. The vision camera understands an object based on its physical attributes, such as shape and size, to effectively control the hardware, which will be used in this work. Besides, the number of objects of a particular section that cross the conveyor to demonstrate the identification of moving objects is counted and displayed. For identifying a good object, the wavelength data is used, while determining how to match the geometric patterns and identify the dimensions, and edge detection is applied. The ability to count specific attributes of objects is tested in different test paths. The sorting of objects using machine vision was performed using an algorithm of pattern matching. A pattern image template was built and stored in a computer's memory. The vision application investigates the image and transfers it to the classifier if the received image matches the model image or not matches. In the conclusion part of the thesis, the results are discussed over the experimental data.