The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. Vol. XXXVII. Part Bl. Beijing 2008 
Figure 5. Algorithm for first method. 
According to algorithm the difference between first and last 
returns must be bigger than a treshold. After tests the treshold 
was chosen 5 (Figure 7). 
Algorithm for intensity drop: 
1. main() { 
2. FILE *fin,*fout; 
3. fin = fopen ("cite2.txt","r"); 
4. fout = fopen ("inten2.txt","a"); 
5. int k; /*N is defined as row number separately, i.e. 
243.400*/ 
6. double Nor[N]; 
7. float Eas[N], Hei[N], Int[N], Easl[N], Norl[N], 
Heil[N], Intl[N]; 
8. if (fm=NULL){ printf("Cannot open the input file"); 
return 1;} 
9. if (fout==NULL) { printf("Cannot open the output 
file"); return 1;} 
10. for (k=l ;k<N+l ;k++) { 
11. fscanf(fm,"%f %lf %f %f %f %f %f %f\n",&Eas[k], 
&N or[k] ,&Hei [k],&Int[k],& Easl[k], &Norl[k], 
&Heil[k], &Intl[k]);} 
12. for (k=l;k<N;k++) { if (Int[k]<35) { if (Int[k+1]<35) 
{if (Int[k+2]<35) 
13. fprintf (fout, "%.2f %.21f %.2f %+.2f %d \n", 
Eas [k] ,Nor[k] ,Hei [k] ,Int[k] ,k); } } } 
14. fclose (fin); 
15. fclose(fout); 
16. getchar(); 
17. return 0; } 
The output is txt file and was imported to las format and 
displayed via LASEdit software. 
Intensity Drop Method: When laser light hits vegetation 
multiple returns used was due to multiple return in trees. 
Figure 6. Due to multiple returns in trees the first and last 
return must have different heights. 
Figure 7. Algorithm for second method. 
Algorithm for height difference: 
1. main(){ 
2. FILE *fm,*fout; 
3. fin = fopen ("cite2.txt","r"); 
4. fout = fopen ("heights2.txt","a"); 
5. int k; /*N is defined as row number separately, i.e. 
243.400*/ 
6. float Eas[N], Nor[N], Hei[N], Int[N], Easl[N], 
Norl[N], Heil[N], Intl[N], df[N]; 
7. if (fm==NULL){printf("Cannot open the input 
file");retum 1;} 
8. if (fout==NULL){printf("Cannot open the output 
file"); return 1;} 
9. for (k=0;k<N;k++) { 
10. fscanf(fm,"%f %f %f %f %f %f %f %fn", 
&Eas[k],&Nor[k],&Hei[k],&Int[k],& 
Easl[k],&Norl[k],&Heil[k],&Intl[k]); 
11. df[k]=Heil[k]-Hei[k]; 
12. if (5<fabs(df[k])) { 
13. fprintf(fout,"%.2f %.2f %.2f %.2f %.2f %d\n",Eas[ 
k],Nor[k],Hei[k],Heil[k],df[k],k+l); } } 
14. fclose (fin); 
15. fclose(fout); 
16. getchar(); 
17. return 0; } 
3. RESULTS 
The first method returned tree points quite satisfactory and 
the second method returned tree points with errors due to 
balcony and bulding comers (Figure 10 and 11) 
4. CONCLUSION 
Even though there are some errors in the second method the 
density of the trees is more satisfied (Figure 12 and 13). 
REFERENCES 
Weitkamp, C., 2005, Lidar: Introduction, in Laser Remote 
Sensing, Chapter 1, pp. 2-36, Eds. Fujii, T. & Fukuchi, T., 
Taylor & Francis Group, Florida. 
429 
Harding, D. J., 2000, NASA’s Goddard Space Flight Center 
Sithole, G., Vosselman, G., 2004. Experimental comparison 
of filter algorithms for bare-Earth extraction from airborne