20.500.12556/RUL-100497 Prepoznavanje oblik in razporejanje sestavnih delov glede na njihove značilke Shape recognition and sorting of parts according to their features V proizvodnji se pogosto srečujemo z razvrščanjem in ločevanjem različnih izdelkov in sestavnih delov. Zaradi monotonosti, hitrosti in potrebe po maksimalni zanesljivosti teh procesov mora biti sistem avtomatiziran. To dosežemo s pomočjo strojnega vida, algoritma za odločanje in manipulatorja za izvajanje premikov izdelkov oz. sestavnih delov. Namen diplomskega dela je razviti sistem algoritmov (strojni vid, krmiljenje robotske roke in odločitveni algoritem), ki prepoznajo oblike delov in glede na postavitev že obstoječih delov določijo pozicijo naslednjega dela. Za preizkus algoritma smo izdelali preizkuševališče s fiksno digitalno kamero za zajemanje pozicije delov in uArm robotom za manipulacijo. Preizkuševališče je posnemalo delovanje igre Križci in krožci, ki vsebuje vse elemente, ki se pojavijo pri razvrščanju delov s strojnim vidom (identifikacija in lokacija delov, odločitveni algoritem in manipulacija). S kamero in algoritmom strojnega vida smo zaznavali objekte in jih ločili glede na velikost njihove površine. Robotska roka uArm je objekte postavila na mesto, ki ga je glede na pravila igre določil odločitveni algoritem. In production we often come across operations as positioning and sorting of assembly parts. Due to monotony, speed and need reliability of this processes the goal is to automate them. This can be achieved by using machine vision, an algorithm and a manipulator that moves the parts and components around. This work's aim is to develop a system of algorithms (machine vision, control of a robotic arm and a decision algorithm) that can recognize certain shapes, and, depending on the position of preplaced shapes, decide the placement of the next one. To test the developed algorithm, we made a testing rig, which consists of a fixated digital camera that captures the positions of parts, and a uArm robotic manipulator for manipulation of detected parts. The system is based on the game known as 'tic-tac-toe', which contains all the elements used in real-life applications of sorting using machine vision (identification and placement of parts, decision algorithm and physical manipulation of parts). The camera and machine vision algorithm were used to recognize objects and distinguish them according to their surface area. The uArm robotic arm was used to place the objects to spots assigned to them by the decision algorithm taking in account the rules of the game. strojni vid krmiljenje robot prepoznavanje oblik odločitveni algoritem machine vision control robot image recognition decision algorithm true false false Slovenski jezik Angleški jezik Diplomsko delo/naloga 2018-03-23 07:45:02 2018-03-23 07:45:07 2022-08-14 03:40:46 0000-00-00 00:00:00 2018 0 0 0000-00-00 NiDoloceno NiDoloceno NiDoloceno 0000-00-00 0000-00-00 0000-00-00 1970-01-01 204384 1042.pdf 1042.pdf 1 A091BD41B18F0CDBC2756D4274D0E565 23f53e31181d83db112ceb057b566f3e88a3e6f37f3faae22fc858019aaefa3e baf3d5e9-a1b4-11eb-a523-00155dcfd717 https://repozitorij.uni-lj.si/Dokument.php?lang=slv&id=110266 Fakulteta za strojništvo 0 0 0