ABSTRACT
Deep learning technology that recognises and classifies photographs in real time on behalf of people requires investigation. Object recognition is challenging when it comes to finding an object of interest in a video clip or photograph. As described in this paper, if the project is successfully implemented, we will be able to easily analyses and even manage the quality of the items being made. In order to decrease human error, obtain 100% error-free products, and maximise income, computers and surveillance are being integrated into the quality control and regulatory process. We can now go even further than standard surveillance cameras thanks to the arrival of computer vision. Thanks to computer vision, we can maintain a standard and quality level that we demand from each and every product that is manufactured. The data set, which is a collection of photographs that reflect the ‘Perfect’ product, can be used to do this. Each unit is assessed and compared to the data set, with the units that closely resemble the photographs in the data set being kept and the defective units being deleted. Experiments demonstrate that it can accurately recognise and classify things in a range of conditions, and that it can track objects in real time because the calculation speed is faster than the previous method.
