A Fingertip Detection Method Based on Kinect Depth Information

Abstract

The invention relates to a fingertip detection method based on Kinect depth information. The device Kinect is connected to a computer through a cable; it is characterized in that the specific steps are as follows: Step 1, extracting hands and obtaining palm coordinates: Step 2, fingertip positioning: including images Preprocessing and hand contour; joint bilateral filtering for the extracted hand area; use the Douglas-Puke algorithm to approximate the specified point set, find out the polygonal fitting curve of the contour and draw the fitting curve of the hand. Step 3. Use the convexHull() function to retrieve the above steps, analyze and obtain the convex hull points of the hand. Step 4. Calculate the curvature of the obtained convex hull point, and set an appropriate threshold to remove the convex hull point at the wrist according to the difference between the curvature of the wrist and the curvature of the fingertip. The method can complete gesture recognition tasks in real time and accurately, improves the real-time performance and accuracy of Kinect gesture recognition, and improves the natural gesture interaction experience.

Description

technical field [0001] The invention relates to a fingertip detection method based on depth information, belonging to the technical field of human-computer interaction. Background technique [0002] With the development of human-computer interaction technology, natural interaction has become the development direction of human-computer interaction technology. In recent years, using human hands for natural and intuitive interaction without wearing auxiliary equipment has gradually become a research hotspot in this field. [0003] The key to realizing natural human-hand interaction lies in the accurate recognition of gestures. At present, gesture recognition methods without wearing assistive devices mainly include methods based on color cameras and methods based on depth sensors. Gesture recognition based on color cameras is easily affected by complex lighting and background conditions, especially when the background is very complex, the recognition accuracy is extremely low....

AI Technical Summary

Problems solved by technology

Many scholars have begun to use Kinect for gesture recognition. However, most of the existing methods have cumbersome steps, complex algorithms, and poor real-time performance.

Images