Abstract
A contemporary study on the Kinect sensor as a visual sensor device for robots shows that the sensor has some fundamental flaws. One of them are shadows on the object edge that will affect the process of recognition of shape (shape recognition) spatially. If the Kinect sensor is used in the robot vision navigation system, the sensor may lead to errors in the robot's decision on the shape of the object sensed by the sensor. The previous research reports a positive influence on the variation of smoothing process by using neighborhood filtering. This research will use multiple neighborhood localized filtering (MNLF) method to eliminate structural noise generated by kinect sensor IR camera. The robot model that will be used for testing is 6WD Wild Thumper Mobile Robot Chassis from Dagu Robotics. The calculation of SSI (Structural Similarity Index) calculation based on ROI between image index 0 (original image) with index 6 (image result after multiple filtering process) results SSI index with value 0.999999930515914. This indicates that multiple filtering processes do not affect the quality of images produced by Kinect sensors. The number of 0.99 can be rounded to 1 so that the conclusion based on ROI image assessment shows no differences on image quality after process.
Highlights
Autonomous machinery has evolved into the area of an artificial intelligence machine lately
Kinect is a vision sensor consisting of a visible light camera and an infrared array sensor [7]
Recent research by Kurniawan reports that the effect of Isolated Neighborhood Averaging Mask (INAM) on the number of pixel noise generated by Kinect has a positive effect of eliminating the noise [3]
Summary
Autonomous machinery (robot) has evolved into the area of an artificial intelligence machine lately. In addition to the algorithm, a set of sensing mechanism is needed to turn the dead machine into a system that is aware of the environment in which it operates Both systems (sensing mechanisms and algorithms) must work in such a way that they can work independently without human control. The Kinect is equipped with an additional sensor in the form of infrared arrays to measure the distance between objects (obstacles) and robot on a certain spatial field. These sensors can provide the robot distance information (depth perception) and simultaneous two-dimensional vision to the robot. The cause-effect relationship between the two (INAM vs Noise) has Pearson coefficient of 0.4, and the saturation layer of Neighborhood-Averaging filter on the pixel noise generated by the Kinect is 2 layer
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
