Recognizing system for three-dimensional object

Abstract

Abstract PURPOSE:To decrease the parameter space memories by using an edge segment as a corresponding unit of a 3-dimensional object to decide 3 degrees of freedom of position and 1 freedom degree of attitude and processing these degrees of freedom in parallel with each other via the general Hough conversion to merge a hypothesis. CONSTITUTION:A computer 2 which executes the picture process of an object model 4 obtained via a TV camera 1 used to acquire an input picture 3 and turns into memory data, a hypothesis generating means 7 produces the hypothesis of the object model with the parameters showing the position and the attitude of the model 4, a hypothesis cluster extracting means 10 merges the parameters in a four-dimensional parameter space 8 into a hypothesis cluster 9 according to the short distances among these parameters, a position/attitude deciding means 11 decides a parameter containing the largest number of hypotheses out of the cluster 9 and then decides the position and the attitude of the model 4 are provided. In such a constitution, the position and the attitude of an object can be decided without securing a memory area for the space 8.

Inventors

• 👤 Ryuichi Tamano
• 👤 玉野　隆一

Bibliographic Details