2004 Fiscal Year Final Research Report Summary
The development of software for measuring motion space
| Project/Area Number |
14540663
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Single-year Grants |
|---|
| Section | 一般 |
|---|
| Research Field |
人類学(含生理人類学)
|
|---|
| Research Institution | Kinki University |
|---|
Principal Investigator |
OHASHI Jun-ya Kinki University, School of humanity-oriented science and engineering, Associate professor, 産業理工学部, 助教授 (90169037)
|
|---|
| Project Period (FY) |
2002 – 2004
|
| Keywords | motion space / video / measurement / software / ergonomics / motion analysis / posture observation |
|---|
| Research Abstract |
The software for measuring motion space with common personal computer, video devices, digital camera and so on was developed. Movie files in AVI format are read, outlines of objects are extracted, the outlines are overlapped and the parts for determining motion space are pointed by operators and the points are transferred to actual scale. The picture without objects, which is called reference picture, is compared with pictures with objects, which are called analysis pictures. The different parts are extracted as objects. The comparisons are applied to plain pictures, color difference (differences between red, green and blue) picture and outline picture with three methods. These pictures are compared in each pixel. In small square, whose side length is between three and nine pixels, correlations are calculated between reference picture, analysis picture and difference picture (difference between reference and analysis pictures). Presence of object is determined with the help of the correlations. These and two other methods of comparison use the reference picture. A color discrimination method is also used for the determination of objects. The method does not use the reference picture. These methods can be chosen and combined for the synthetic discrimination. Plural objects and restricted parts of whole objects can be used with the color discrimination method. Motion and posture can be also analyzed as the influence factors of the motion space with the increase of freedom of the objects selection and calculation of center of each object. Information of presence of object at each pixel can be outputted as distribution pictures and numeric data files. The information is frequency alternation times, maximal presence duration, maximal absent duration and standard deviation. Motion space can be determined from the quantitative space use with the information.
|
