Kernel Density Estimation from a 3D point pattern in

Hi,

I am interested in spatial statistical analysis at a beginner user level. I am trying to find an existing software that can perform kernel density estimation on a 3d spatial point pattern within a 3d window with "no go" zones, ie regions where no points exist. I have recently found a package called ks in R that does multidimensional kernel smoothing, but I am not able to introduce no go zones. Could someone guide me to another code or a possible mathematical work around for this? 


Thanks in advance

Vijay