Face Recognition with OpenMV Module for Adjusting Modern Hand Assembly Place

V sklopu Industrije 4.0 predstavlja pomemben in dostikrat spregledan segment tudi ergonomija in aktivno prilagajanje ročnega montažnega mesta delavcu. Za prilagoditev ročnega montažnega mesta je potrebno delavca prepoznati, kar se trenutno največkrat izvaja s pomočjo RFID kartice ali vpisom kode delavca. Vendar pa imajo ti sistemi številne pomanjkljivosti kot so npr. izguba časa, stalno nošenje RFID kartice ipd. Zato smo v zaključni nalogi razvili sistem za prepoznavanje obrazov, ki ga bomo v prihodnosti uporabili za prilagajanje ročnega montažnega mesta v pametni tovarni, ki je glavna ideja industrije 4.0. Na začetku smo predstavili osnove Industrije 4.0, ergonomijo ročnega montažnega mesta, RFID in ostale načine identifikacije, ki so potrebni za razumevanje koncepta 4.0. Nato je sledil razvoj algoritma za prepoznavanje obrazov z OpenMV modulom. Z uporabo OpenCV knjižnic smo razvili program, ki primerja zajeto sliko z interno bazo obrazov. Glede na ujemanje izbere najboljši približek in izpiše ime zaznane osebe. Na koncu smo izvedli analizo razvitega algoritma s petimi različnimi osebami in petimi ponovitvami, s čimer smo potrdili delovanje sistema.Within Industry 4.0, ergonomics and active adjusting of the manual assembly place are very important, but quite often overlooked segments. To adjust the manual assembly place, the worker needs to be identified, which is most often done with an RFID card or by entering of a worker’s ID code. However, these systems have shortcomings such as time wasting, always having an RFID card present etc. That is why in the final task, we developed a facial recognition system that will be used in the future to adjust the manual assembly place in the smart factory, which is the main idea of an industry 4.0. At the beginning of the final paper we presented the basics of Industry 4.0, the ergonomics of the manual assembly place, RFID and a few other systems for face recognition that are important for understanding the 4.0 concept. Then we developed an algorithm for face recognition using the OpenMV module. By using OpenCV libraries, we developed a program that compares captured photos with an internal data base. Depending on the matching results, it selects the best approximation and displays the name of the chosen person. In the end, we tested the developed algorithm with five different individuals and five repetitions, which confirmed the functioning of the system