# How to remove redundant points

We basically have points table including vehicle routes.In huge traffic conditions,devices collecting data on route may output redundant points where at any coordinate same points may repeat.We can easily search and remove those points,but we have another issue like this.

Since points have threshold like 5 meters (for example) among them,when in traffic output may occur so many close(adjacent) points that we cannot seem to analyze.Also we cant adjust device,we have to move on with what we have.

We need to implement an algorithm or script to remove these redundant points.Or at least we can join them into one as we calculate some fields using functions like avg,sum.

How can we achieve this ?
NOTE: We have coordinates in mapinfo table but as points we can move them into anything,your answers does not have to be in one platform.

Thanks

EDIT:Open source solutions would be ideal.

-

In PostGIS I have searched for proximity points ( in a meters tolerance ) and consider points inside this area are redundant. The code is like this :

``````SELECT * from (
SELECT u1.serial, ST_Distance_Sphere(u1.the_geom, u2.the_geom)  as distance, u2.serial
FROM units u1, units u2 where u1.serial <> u2.serial order by distance asc ) as t1
WHERE t1.distance < 200
``````

I have a "UNITS" table ( points of geometry in the_geom column ). For each point, I search for points in a given distance ( less than 200 meters ).

You can fix the u1.serial and do a individual search.

-
I don't have PostGIS,but the query may not differ in other sql solutions.I'll take a closer look at the query.Thanks. –  Myra Dec 3 '12 at 7:51
Well, ST_Distance_Sphere makes use of PostGIS extension. Maybe you will need to have. Standards SQL will not solve your problem or will give you a lot of pain. –  Magno C Dec 3 '12 at 10:40

The Integrate tool in ArcGIS will do what you're asking. The help file shows boundaries, but it works on points, too. You'll have to experiment with the tolerance value and inspect the output to determine what value is best for your data.

-
+1 Integrate would be ideal –  Aaron Nov 30 '12 at 16:38
Does Integrate tool also make aggregate functions while copying features ? –  Myra Dec 3 '12 at 7:35