2. Aruco Markers and Dictionaries

okay so first of all let's talk about markers okay detection of square markers like the one that the ruko use is something that has been on the literature for many years so basically different researcher different developers have proposed their type of markers in essence all of these these are called square flaner markers because you know the r-squared your plan okay they're easy to print and every useful because you can estimate the post of a camera by just using the four corners if you detective so different there are different libraries for doing this you have our tool kit you have our tire chile tag you have many different ones normally each developer of the library have created their own set of markers because you don't only want to have one marker but you have normally you may wanna have different markers maybe they each represent a different object in our materiality application or whatever the thing is that in essence this marker are always the same type you have basically an external black border which is black okay this is what we see here hopefully we see that here okay okay and in here okay there are a set of grid here of beads black and white okay so the way these bits are filled with black and white are a binary code that represent the identification of each marker okay so you see that you can become optionally you can obviously see that you can have different configurations you could have markers where this grid is three times three or four time four or five time five you know many different approaches so basically how it is impact in the detection well if the less bit you have the bigger I mean you can detect this marker out of larger distance but you have less possible markers to represent if you increase the number of beats of course you can represent many more markers but if you go far from the camera then this beats will be very small and then you will not be able to attack them from so this is something that has been a studied and there is a we are working with it is the scribe methods to optimally create this set of markers okay by the way a set of marker is what is called dictionary a dictionary of markers okay well the good thing is that Arouca the library can detect most of the of the dictionaries that have been previously proposed okay it can detect average tags to tags are to keep dictionaries so you can use a leuco to detect all these dictionaries but also in one of our papers we develop method to generate optimal dictionaries and so there is one dictionary which is called meep thirty-six I will put the link here down okay down here and this is the one we have developed and this is an optimal marker so we recommend to use this in all your one thing about markers you create in print and that I will show you how to do later is that it is important to remark that in order to properly protect marker the marker should have white space here around black border okay the detection algorithm this to be person here okay if you create a marker and you cut it and you don't leave whitespace okay and if you try to put markers like this okay what happened the detection is not going to be okay okay detection will fail you need to leave an appropriate white border around the black border for a proper detection okay no thank you explain how to do the first step which is to get unpretty markers okay you can go to the SourceForge webpage and go to the local library and then you will have their files and as I told you before there is a recommended dictionary which is the one you have here so if you click you to unload it okay okay wait downloads here you have a bunch of this is the set of markers of the dictionary okay so if you open it okay here's a image you can copy a paste on your I know any domain writer and then print it however you want okay so take a look here this number here is the ID of the marker okay it's different marker here's as a different ID which is here represented in a light grey color so when you print it you know which one is it okay so the first thing you can do you download this and you try to bring this on pieces of paper okay the second step is you to download the library okay if you go into the the REO 3d library folder okay what you find here is this is a source code that we compile both Lionel's Windows and Mac okay and if you want if you don't want compile because you don't know how to compile but you just want to use a library okay you can download here for window to binaries okay so it will download the virus here I'm gonna explain you how can you do EC detection okay this downloaded let allowed in the library okay here it is in here what you will see is this is like a set of folders mainly prepared for developers so they can easily put this on their project but if you are not a developer you have to go to this place here and you can find a set of programs okay these are the tools of the libraries that you can use even if you don't if you're not a programmer okay okay so I now wanna focus on iroko symbol this is the first test that you should read if you are a developer you can go to source code we will go into the tailor but if you want to use it you go to our simple and you execute this program I'm gonna do it in I'm gonna do it here okay so this is a low cost simple if you run it it will say it will tell you some which are the arguments you see here first you have to provide either an image or a video showing one marker then this we will forget for a while okay so what are we gonna do okay we can do the following we take off we take our mobile phone or any camera we have and we record video yes showing like sorry yeah I'm gonna put the marker here I'm gonna record the video okay okay let's record the video okay I'm recording here so okay okay I have already downloaded the video to my computer and we're gonna run okay these Perman to see what happen when we run the program so we basically say the name of the file here and then we press into what I wanna see here is this okay we are going to see here but a bunch of information about the detection times and so on maybe you don't see this in the developer version but it's the bad version and also here you see the marker that has been detected and the four-corners in the image okay okay so this is the number you see here in the image sorry this is the detected marker okay and this is the number of the marker this is the identification of the market if you look closely up at the marker you will see that it is printed here the number 61 probably see him sing here but it's it's here okay so you see in red now you see the sides of the marker and in blue number here it's the ID of the marker okay and okay so we have already seen our first detection compute elations

Channel: Rafael Muñoz Salinas

Views: 26,657

Keywords:

Id: aWmzhmfahh8

Length: 10min 36sec (636 seconds)

Published: Tue Jan 30 2018