Project # 12 – Smart Walker REU student: Christopher Feltner Graduate mentors: Sharare Zethabian and Siavash Khodadadeh Faculty mentor(s): Dr. Turgut and Dr. Bölöni Week 4 (June 4 – June 8, 2018) Accomplishments: Created a function to average depth values across the center of an image while negating error state values, creating a less noisy 1-D NumPy array of distance values Created a function to create a piecewise linear segmentation of distance values in order to differentiate regions of floor space from obstacle space Developed a method to identify obstacle distance Problems & Solutions: Had an issue where accumulated error over large regions of curvature would improperly indicate segmentation Solution: Added a “window” to segmentation function which only considers points in a given range in calculations of standard error Areas of floor space are not represented linearly. They are actually represented as a quadratic. This is a quirk of the Kinect camera Algorithm sometimes has issues when an object is directly in front of the camera and is very close Plans for next week: Fix distance function Identify a solution to the problem of nonlinearity in floor space representation Testing! NSF REU Research Experience on Internet of Things 2018
Project # 12 – Smart Walker REU student: Christopher Feltner Graduate mentors: Sharare Zethabian and Siavash Khodadadeh Faculty mentor(s): Dr. Turgut and Dr. Bölöni Figure 1: A depth image with the average function being applied NSF REU Research Experience on Internet of Things 2018
Project # 12 – Smart Walker REU student: Christopher Feltner Graduate mentors: Sharare Zethabian and Siavash Khodadadeh Faculty mentor(s): Dr. Turgut and Dr. Bölöni Index 0 Distance value (mm) Array index (top is 0) Index 480 Figure 2: The plotted values of the array of averaged depth image values (left) and the original depth image with indices shown (right) NSF REU Research Experience on Internet of Things 2018
Figure 3: Plot with segmentation shown Project # 12 – Smart Walker REU student: Christopher Feltner Graduate mentors: Sharare Zethabian and Siavash Khodadadeh Faculty mentor(s): Dr. Turgut and Dr. Bölöni Distance value (mm) Array index (top is 0) Figure 3: Plot with segmentation shown NSF REU Research Experience on Internet of Things 2018
Project # 12 – Smart Walker REU student: Christopher Feltner Graduate mentors: Sharare Zethabian and Siavash Khodadadeh Faculty mentor(s): Dr. Turgut and Dr. Bölöni Distance value (mm) Array index (top is 0) Figure 3: Plot with segmentation shown and regions of interest highlighted NSF REU Research Experience on Internet of Things 2018
Project # 12 – Smart Walker REU student: Christopher Feltner Graduate mentors: Sharare Zethabian and Siavash Khodadadeh Faculty mentor(s): Dr. Turgut and Dr. Bölöni Figure 3: Depth image of obstacle (left) and its segmented center array graphed (right) NSF REU Research Experience on Internet of Things 2018
Figure 3: Graph with areas of interest highlighted Project # 12 – Smart Walker REU student: Christopher Feltner Graduate mentors: Sharare Zethabian and Siavash Khodadadeh Faculty mentor(s): Dr. Turgut and Dr. Bölöni Closest point Figure 3: Graph with areas of interest highlighted NSF REU Research Experience on Internet of Things 2018