Announcement

**Mike Kocourek** · December 17, 2014, 01:49 PM

Re: please help, Vicon v.s Motion analysis corp

Originally posted by boyce60225 View Post

Hi, all
My lab wants to buy a set of camera system to capture the motion of robot. So far, there are two companies (Vicon and Motion Analysis Corp.) gave us full set of motion analysis system.
Vicon offers us T20s camera*5, and Motion Analysis Corp. offers us Raptor-E camera*6. the first offer is slightly more expensive than second one. Though T20s seems to have better spec.(690fps, 2megapixels), Raptor-E(500fps, 1.3megapixels) has more cameras and relative low price. Since I have no experience on these system, there are several things confuse me. Can anyone help us with the following questions or give us any suggestion?

The purpose for us to buy the motion analysis system is that we want to track the motion of robots, like RHex(6 feet), Qudruped(4 feet), etc.. The work space is approximate to 6 meters square. We want to track not only the rigid body part but also flexible leg's motion.

1. Though both system claim that the spatial resolution is less than 1mm, we don't know what condition the system takes can promise that. We should get poor resolution when increasing the distance, right?
2. The other problem is that sometimes we might have to conduct experiment outdoor, so the light condition is unpredictable. Is it possible for camera completely reject the noise generated from the sun light?
3. Since the robot's leg might shade the body part. Are 5 T20s camera enough to handle the shading problem?
4. Motion Analysis Corp. claims that Raptor-E can increase shutter speed up to 10000fps. Dose anyone know how the performance changes in high speed mode?

p.s. sorry for my poor English, but I really need some advices and answer
謝謝(Thanks)

National Taiwan University, Bio-inspired Robotics Laboratory

http://140.112.46.47

Po-Wei, Tseng

Hi,

I wanted to provide some additional assistance to help compare the two companies offering.

The ability to track the 3D motion of various robots can be achieved successfully with the right combination of motion capture cameras and software.

Additional cameras provide better line of sight coverage to track the flexible leg’s motion along with the other rigid body parts in the capture volume desired.

The 6 camera Raptor-E option would provide more coverage to obtain the robust and accurate data that you are looking for.

I have sold systems for both Vicon (16 years) and Motion Analysis Corp. (2 years) and helped hundreds of users in North America with their motion capture needs.

Here are some answers to your questions.

1) The spatial resolution of a motion capture system combines a number of factors including camera design, software features, distance from cameras, quantity of cameras, position of cameras, calibration, and various other factors. The Raptor-E cameras come with zoom lenses which allow the users to optimize the capture volume while adding flexibility to where the cameras are placed. The Raptor-E cameras have real time software assisted focusing to ensure consistent optimal focus with no trial and error. The Cortex software, that is used with the Motion Analysis Corp. cameras, has live 3D calibration quality feedback and the ability to refine the calibration in one click if someone bumps a tripod or camera movement occurs after the calibration.
2) MAC Raptor motion capture cameras have the ability to independently adjust the strobe intensity, exposure duration, background noise, and frame rate. These unique functions combined with special real time on-board processing algorithms allow the user to easily optimize the system in different lighting conditions and accommodate various speed of movements. Many users around the world have used our system to capture data in varying lighting conditions outside, even in direct sunlight on the snow. The unique functions on the Raptor cameras allow them to eliminate any detrimental effects from reflections and detect markers much further than competitors cameras in similar conditions.
3) There will be more occlusions or gaps in the data when using 5 cameras so the robustness of the data will be sacrificed. The 6 camera Raptor E system would provide more line-of-sight coverage and improved robustness, fewer gaps and more accurate data.
4) The ability to precisely configure the camera settings, along with strobe intensity and background noise reduction allows the user to obtain optimum performance at any speed of capture. The competitors’, cameras without these capabilities, will have reduced performance at higher speeds.

Just as they have found that the quality of photographs from digital cameras is not merely dependent upon the resolution of the imaging chip, the overall performance of a 3D system involves many factors. Therefore a set of Raptor-E cameras with 1.3mp can outperform a set of 2+mp cameras from a competitor because of the superior on-board processing, calibration, 3D reconstruction algorithms, marker identification, and other features of the system.

Feel free to contact me with other questions or help with comparison details.

Best Regards,

Mike Kocourek
Vice President of Sales
Motion Analysis Corporation
3617 Westwind Boulevard
Santa Rosa, California 95403
D: 707-579-6520
E: mkocourek@motionanalysis.com
www.motionanalysis.com

**Lasse Roren** · December 18, 2014, 05:05 PM

Re: please help, Vicon v.s Motion analysis corp

Hello,

Having worked with the Vicon system for almost 18 years now, and having observed competitive systems for a similar amount of time, I would like to add some comments.

1) Spatial resolution of less than 1mm is certainly no problem - in most cases spatial resolution will be better (and sometimes much better) than this. It is true that the spatial resolution will decrease as the distance from the cameras increase. The key to spatial resolution (and to minimize measurement error) is the ability of the camera to accurately calculate where the centre of a marker's image is in 2D. Since the technology of all passive optical motion capture system is based on emitting strobed light from the camera and then to record the light that gets reflected from the marker, the main difference between different technologies will be how advanced the treatment of the reflected and recorded marker image is. In the "ideal" world, the camera emits light and the marker is recorded on the camera's sensor as a perfectly circular image. Unfortunately, most of the time the world is not ideal and there are several sources of error - for example, when the images from two marker partially overlap or when the image is partially occluded. In this case, calculating where the centre of the marker actually is becomes much more difficult - imagine that you only see half a marker, and that this marker is 10mm in diameter - if you simply calculate the centre of this, your calculated centre will have an error of 2-3mm. Vicon introduced the concept of recording the marker in full grey scale about 10 years ago. Other manufacturers look at the marker image and simply apply a threshold - if a pixel is bright enough then it's declared to be from a marker, but if it's below the threshold then it's not a marker. This is OK for the "ideal" world where the markers are perfectly circular and where there's no measurement noise. However, in the real world this is most often not the case, and Vicon's greyscale recording means that advanced algorithms that use especially the greyscale information from the edge of the marker image to more accurately find the centre of the marker. This has two major benefits - first, the centre of a marker image can be calculated with higher 2D accuracy, which results in higher 3D accuracy when images from more cameras are combined. Second, the difficult images where markers are overlapping or partially occluded are not thrown away at the camera, but sent to the PC where advanced algorithms will be used to resolve the images. This is essential if you're going to claim spatial resolution of less than 1 mm in real-world scenarios. I am not aware of a single instance where Vicon's been beaten on accuracy and spatial resolution in a side-by-side comparison, and I believe a big contributing factor to this is the way the system deals with the raw greyscale data.

2) Outdoor capture is clearly something all the main manufacturers claim they can do. Vicon's system has been successfully used to capture, among other things, kangaroos jumping around under the hot Australian sun - check out https://www.youtube.com/watch?v=KKyHVe-k9Yg for a nice video.

3) There is, of course, an inherent benefit of having an extra camera since this gives you another angle from which to measure. However, the system's ability to use all the information from the cameras you have is also important, and I believe Vicon has an advantage here. For example, the T20 gives you a 67 degree field-of-view, which is much more than you would get from an 18-55mm zoom lens. This again means that if you're using the 18-55mm zoom lens you either have to move the camera much further away, or you have to live with a much smaller measurement volume.

4) Increasing the frame rate has two main impacts on a motion capture camera (or indeed any camera): 1) the shutter speed has to be shorter, which means that less light will be let onto the sensor, which means that the camera will have less range, and 2) the area of the sensor which actually records data goes down because there's simply not enough time between each frame to read the whole sensor. For the Vicon T20, the on-board processor is fast enough to read the entire 2 mega-pixels until 690Hz, whereas the Raptor-E can, according to the specs on the web site, only read the entire 1.3 mega-pixels until 500Hz. This has potentially a big impact if you go up to 1000Hz - the T20 will still give you more than 1 mega-pixel worth of usable data, whereas other cameras will give you much less.

I hope this helps!

Best regards,

Announcement

please help, Vicon v.s Motion analysis corp

please help, Vicon v.s Motion analysis corp

Comment

Comment