| Name: | Description: | Size: | Format: | |
|---|---|---|---|---|
| 14.02 MB | Adobe PDF |
Advisor(s)
Abstract(s)
A necessidade crescente de obtenção de melhores imagens, leva a criação de lentes mais complexas, o que obriga a que existam mais cuidados na sua produção, de forma a evitar aberrações nas imagens. Consciente desses desafios a empresa Leica identifica lacunas no seu processo que visa minimizar com esta dissertação, colocando ao dispor os seus melhores profissionais na validação no processo desenvolvido. A qualidade de uma lente depende do processo de alinhamento entre as várias superfícies óticas e a superfície exterior que a reveste. Atualmente, o processo de alinhamento de todos os elementos que constituem uma lente encontra-se ultrapassado, apresentando dificuldades em responder aos crescentes desafios óticos e mecânicos, para além de requerer a presença de um operador experiente para efetuar esse alinhamento. O desenvolvimento de ferramentas tecnológicas permitirá aumentar a sua eficiência e libertar o operador de trabalhados repetitivos prejudiciais `a sua saúde. Na dissertação propomos o desenvolvimento de um sistema de suporte ao processo manual ao criar um sistema de visão que ira supervisionar e ajudar o operador no processo de alinhamento de lentes. E desenvolvido um algoritmo de identificação de desvios das várias superfícies óticas que constituem uma lente. Com base em técnicas de processamento de imagem e de Inteligência Artificial (AI), são desenvolvidos algoritmos de previsão para estimar as posições que correspondam ao alinhamento de todas as superfícies óticas. A partir desta previsão será possível quantificar rigorosamente o sentido e a grandeza dos desvios, e assim fornecer ordens precisas para o sistema de alinhamento.
The need to have better images, had consequences on the development of the lenses making them more complex. The Leica company is aware of these challenges when identifying gaps in its process that it aims to minimize with this dissertation, providing the best professionals in the validation of the developed process. The lens quality depends on alignment process between each optical surface and between the outer surface of the lens. Currently, the process used in the company is already obsolete and presents great difficulties to give a great answer to the growing optical and mechanical challenges. In addition, this machine must have an experienced worker working to do a correct alignment of all lens parts. The development of technological tools will allow you to increase your efficiency and free the operator from repetitive work harmful to your health. In the dissertation it is proposed the development of a support system for the manual process by creating a vision system that will supervise and assist the operator in the lens alignment process. An algorithm was developed to identify deviations from optical surfaces in the lenses. Based on image processing techniques and Artificial Intelligence (AI), prediction algorithms are developed to find the alignment of all optical surfaces. From this forecast, it will be possible to accurately quantify the direction and magnitude of the deviations, and thus provide precise orders for the alignment system.
The need to have better images, had consequences on the development of the lenses making them more complex. The Leica company is aware of these challenges when identifying gaps in its process that it aims to minimize with this dissertation, providing the best professionals in the validation of the developed process. The lens quality depends on alignment process between each optical surface and between the outer surface of the lens. Currently, the process used in the company is already obsolete and presents great difficulties to give a great answer to the growing optical and mechanical challenges. In addition, this machine must have an experienced worker working to do a correct alignment of all lens parts. The development of technological tools will allow you to increase your efficiency and free the operator from repetitive work harmful to your health. In the dissertation it is proposed the development of a support system for the manual process by creating a vision system that will supervise and assist the operator in the lens alignment process. An algorithm was developed to identify deviations from optical surfaces in the lenses. Based on image processing techniques and Artificial Intelligence (AI), prediction algorithms are developed to find the alignment of all optical surfaces. From this forecast, it will be possible to accurately quantify the direction and magnitude of the deviations, and thus provide precise orders for the alignment system.
Description
Keywords
Lens alignment Artificial Intelligence (AI) Prediction algorithms