LOT 
31: 
on, 
in 
L7 
), 
Ih 
18Y 
)n. 
111 
iper. 
[lu- 
the 
)ded 
reat 
-- Weighted gravity center algorithms . 
-- Gradient Parabolic Interpolation 
-- Sample Moment Match 
-- Least Squares Straight-line Fitting 
The test results show that , for location mark of 
the control point using the mark pattern match 
and least squares straight-line fitting algori- 
thms, the precision of location is higher, for lo- 
cation mark of the feature point on the moving 
body using the weighted gravity center algori- 
thms the precision of location is satisfactory 
and the computing is easy and stable . 
  
2D MASK EDGE SUB- MARK 
MATCH MATCH PIXEL | | EDGE | | CENTRE 
CRUDE | DETECT [3] EDGE [4 FITT [-1 LOCATE 
LOCATE DETECT 
  
  
  
  
  
  
  
  
  
  
  
Fig.8 Mark Point Detected Procedure 
4.9 Objective Points Positioning 
The equation of 3D-DLT (Direct linear transfor- 
mation ) can be written as : 
x+k1» (x-x0) #r2- (L1+X+L2»Y+L3*Z+L4) 
/ (L9*X+L10*Y+L11*Z+1) = 0 
y:kl* (y- y0) *r2- (L6#X+L6+Y+L7+Z+L8) 
/ (L9*X+L10*Y+L11*Z+1) = 0 
if all signing points are the same plane , which 
vertical with the light axis of the CCD camara, 
the equation can be transformeted into 2D-DLT 
equation: 
x+k1* (x-x0) *r2- (L1+X+L2+Y+L3) / (L7+X+L8*Y+1) =0 
y*kl* (y-y0) *r2- (L4eX+L6eY+L6) / (LT+X+L8+Y+1) =0 
6. TEST AND CONCLUSIONS 
The some primitives tesis have been performed 
for study human movement, as shown in the Fig. 9 
and Fig. 10.0n Fig. 9 some trajectories of feature 
points on a moving human is exhibited. Here (a), 
(b and (c) respectively is walking, running and 
jumping. On Fig. 10 some analysis results is shown 
‚here (a) is the stick diagram when a human run- 
ning, (b) is a angle-time picture of knee, (c) is 
a angle-time picture of ankle. 
By adopting this system and its moving object 
surveying method, the sub-realtime measurement of 
the moving body can be realized It can also per- 
form the description of the loci, velocity and 
acceleration of the movement of the body. The 
  
   
      
   
  
  
  
  
  
  
  
  
  
  
   
   
   
   
   
   
   
    
  
  
   
  
  
   
    
   
  
   
  
  
  
  
  
  
  
  
  
   
  
  
     
initial experiment on the human gait measurement 
has indicated that this system can fulfil the 
requirement of medical analysis . 
6 . REFERENCE 
Wang Zhizhuo , Principles of photogrammetry (with 
remote sensing).Press of WISM. Publishing House 
of Surveying and Mapping Beijing 1990 
Lin Zongjian ,Lu Jian ,Fu Zhongliang . GIS inpu- 
ting through CCD camera , PROGRESS IN DATA ANA- 
LYSIS Commision 3 of ISPRS . 1990. 
Lin Zongjian ,Lu Jian et al , A topographic con- 
tour lines reading system . Proceedings of the 
11th ASIAN CONFERENCE ON REMOTE SENSING. 1990. 
Abdel Aziz, Ÿ. I. and Karara, h.m., 1971. Direct Linear 
Transformation from Comparator Coordinates into 
object space coordinates in close-range photogra- 
mmetry. Proc. ASP/UI Symp. on Close Range Photogra- 
mmetry, Urbana, IL. pp. 1-18 
Woliring H.J. and Marsolais, E, B., 1980. Optoelectric 
(SELPOT) gait measurement in two and three dimen- 
sional space . Apreliminary report. Bull.Prosth. 
Res. , 17: 46-62. 
Jarrett, M. 0, , Andrews, B. J. and Paul, J. P., 1976. À te- 
levision computer system for the analysis of ki- 
nematics of human locamotion. IERE conf. Proc. , 34: 
367-370, 
Ferrigo,G. and Pedotti,A., 1986. Elite: A digital de- 
dicated hardware system for movement analysis 
via rael-time TV signal processing. IEEE Trans. 
Biomed. Eng. (BME), 32: 943-949. 
Mitchelson, D. L. , 1975. Recording of movement without 
photograph. In: Technigues for the Analysis of Hu- 
man Movement . Lepus Books, London.