We recorded videos showing a spinning pinwheel. Many thanks to Ma'am Jing for letting us borrow her camera. We then made the pinwheel spin by walking and running while holding it. These two instances were recorded with a camera of certain fps (frames per second). One of the videos were shown below.
Figure 1: Running with the pinwheel
The videos were parsed to obtain a set of pictures that will be color segmented. Only a part of the videos were considered; in particular, one spin of the pinwheel. Notice that there are different colored papers on the the pinwheel. These colored papers will be the region of interest (ROI) that will be used for parametric segmentation. In particular, we used the reddish paper as the ROI. Below is one of the segmented results of the parsed images from Figure 1.
Figure 2: A Segmented Image
Considering a point from the patch like in Figure 2 and considering all of the segmented images, one can trace out the spinning movement of the pinwheel. Below are the traced movements of the colored paper on the pinwheel for the walking and running instances.
Figure 3: Path of the colored paper while running
Figure 4: Path of the colored paper while walking
One may notice that there are more data points (parsed images) for the walking instance since it will take longer for the pinwheel to complete one spin in that instance. The shapes are circular which is expected as the spinning pinwheel should trace out a circle. Some data points for the walking instance were out of the circular trend (not shown here) and can be taken into account from the choosing of a point from the patch of the segmented image.
No comments:
Post a Comment