<div dir="ltr">The code in kdtree.hpp includes this comment:<div><br></div><div><div> // NOTE: this will visit points based on 'Manhattan distance' aka city-block distance</div><div> // aka taxicab metric. Meaning it will find all points within:</div>
<div> // max(x_dist,max(y_dist,z_dist));</div><div> // AND NOT than what you would expect: sqrt(x_dist*x_dist + y_dist*y_dist + z_dist*z_dist)</div><div> //</div><div> // This is because it converts the distance into a bounding-box 'region' and compares</div>
<div> // against that.</div><div> //</div><div> // If you want the sqrt() behaviour, ask on the mailing list for different options.</div></div><div><br></div><div>I could implement this, but thought I'd ask if this is a solved problem.</div>
<div><br></div><div>Thanks!</div><div>Tom</div><div><br></div></div>