this is brilliant shezan, although i have no idea how to import objects to visitor, I've tried importing the .exp file using the import objects script, it processes for a short time but then it rejects all the objects.
Your procedure was formally correct.
Rejection was due to the fact you must have the objects already loaded on Visitor 3 "Nature object definitions" and their NAMES must match with world tools "item type name"
In the above image:
1) open Nature Objects definitions in V3, add objects by getting them from your "ca" folder or other custom ones
2) the program automatically will give the current path and file name to the "object name".
You can change the object name as you like but you must be sure that this object name MATCHES (3) with the Item type name set on Object Management window in World Tools.
The Name must be freely customized for clarity purposes in V3, but this name must match the imported one (World Tools uses Item Type Name field to export objects).
If imported file doesn't have a reference in the Nature object definition names or Artificial object definition names the import script will fail (reject) all unkown objects.
Hope this was clear, if not i'll be glad to explain you again
Ok, here I used the forest generator on a larger map (10km x 10km). I used a sat image from Google earth (reduced the number of colors down to 4) as mask-image for the generator. The area I choose has a lot of trees, so in the end Worldtool generated 1.6 mio single trees (wow).
Great
I wonder how performance goes with 1.6 millions trees on a single map.
In case you have placed too much trees you can reduce them by acting on maximum objects per pixel and placement probability on single colors.
did you have an ingame screenshot of your results? will be very interesting!
Your tool looks like a big improvement from MaskMapper, but could you tell me how you manage to get some randomization in tree placement so we can't get the vegetation grouping we have previously and also get less trees if needed.
@jens198 : The high tree number can be a real issue ... that why I have never finished my 1st Sbrodj ArmA map , more than a million trees and a beautiful but unplayable map ... . What is great here is that you can rebuild a vegetation cover on the spot ... not being depressed by 15 working days loss.
yep worked a charm, thanks Shezan, i pretty much got good results strait away and it was very easy to use, now ill replace all my vegetation using this tool, thanks again.
To reduce number of placed object you must work on the "object per pixel" number that means the MAXIMUM ALLOWABLE OBJECTS that COULD BE PLACED in a pixel. If you have a world size of 10240 and a map of 512x512 each pixel is a 20x20 meters area. Setting 1 to this parameter is maybe a small value for a forest.
The second parameter is the COLOR PROBABILITY. If your area is too much populated by objects you must reduce color probability on this area.
The right combination between maximum allowable objects and color probability will help you find the correct mix between mask image size and object placement density.
How it works: (for dummies, very long )
Spoiler:
Well, when you load a mask the program will check image mask size with the world size you need in your project (and you have defined at project creation in WT).
Based on the mask size and the world size you have an equivalent area represented by your pixel.
Example: your map is 10240x10240 and your mask picture is 1024x1024.
Each pixel of your map represent a box of 10 meters x 10 meters on the map.
for each color in the mask you can set the "maximum object per pixel" value.
If this value is set to 1 the program will ATTEMPT to place 1 object per pixel, that means 1 object (a tree, for example) on each 10x10 mt square (in a random place inside the box).
Setting a number bigger than 1 means the program, for each pixel, will ATTEMPT to place more than 1 object (by default a random number between 1 and the number you set there or, if option behind is unckecked, the max number itself)
Beware, the program will ATTEMPT to place the object...
So, this is the workflow:
the program will find the pixel, and calculate the relative coordinates represented by the pixel (in our example, the square coordinates for a box of 10x10 meters).
The program will determine the COLOR of the pixel and get the color configuration you choosed.
If you have set a MAXIMUM OBJECT PER PIXEL = 5 on this color the program will:
Find a random number between 1 and 5. Example: 4
For 4 times the program, in the selected pixel area will launch a random number between 1 and 100.
If random number is < than color probability (70% in our example) AN UNDEFINED object is SUCCESSFULLY PLACED.
But... what object? the object is again randomly selected between objects indicated in the selected color. The PROBABILITY set on the object determines the probability than this object will appear (note, again, an object MUST appear if color probability determines that an object is successfully placed, object probability is for choice BETWEEN objects).
Let's follow the program flow:
program will find color GREEN.
Green probability is 70%
Objects behind green color is:
Pinus 80% (1-80)
Small rock 10% (81-90)
Small bush 10% (91-100)
Max number of objects on Green color is 5
Random number betweek 1 and 5: Result = 4
First try (1/4)
launch random number: 87
87 > 70% probability on this color.
Launch fail.
Second try (2/4)
launch random number: 32
32 < 70% probability on this color.
Placement OK
Object selection: random number 56 (1-80 is pinus)
Pinus selected
randomize position in the pixel area
place pinus.
Third try (3/4)
launch random number: 66
66 < 70%. Placement ok.
Object selection: random number: 83 (81-90 is rock in our example)
Small rock selected
randomize position in the pixel area
place rock.
Fourth try (4/4)
blah blah blah
MOVE AHEAD TO ANOTHER PIXEL
restart procedure
---------- Post added at 07:39 PM ---------- Previous post was at 07:30 PM ----------
Originally Posted by Old Bear
so we can't get the vegetation grouping we have previously and also get less trees if needed.
About grouping: define a color for a vegetation type:
for example:
green = a pinus forest
objects behind green color will be pinus and maybe some rocks.
light green = a betula forest
objects behind light green color will be betula and betula bushes, for example.
red = rocks
objects behind could be rocks, with a very low probability of appearance for COLOR PROBABILITY (10%?)
blue = grass field
objects behind could be sunflowers or other kind of fields, with very high density x sq/mt (high "max object placed per pixel" value) and high "hit" probability on the color itself.
Mixing your mask with your color elevation DEM will allow you to have a fine tuning on your object placement densities... a degradation from light green to dark green can be used, for example, for gradually changing forest density from high to low, and contaminating a betula forest with pinus when the height increases... There are 256 possible color combinations that could be used in the way you like
Let me know if you need more help
---------- Post added at 07:48 PM ---------- Previous post was at 07:39 PM ----------
Finally, this is my mask (reduced to 512x512 for showing purposes).
It's the merge between altitude mask and forest / vegetation mask
This image is not usable for my needs due to HIGH number of color variation (256) that will make the work too hard for me and will be reduced to 32 colors or so.
Very very very impressed, this just saved me days on a small map I am building!
CBF Ship Building Industry
Former leader of the now closed 1982: Flashpoint in the Falklands team Permissions still required to use FLK models, all it takes is asking (via pm not a thread please)
HOME 
Originally Posted by jens198
Ok, here I used the forest generator on a larger map (10km x 10km). I used a sat image from Google earth (reduced the number of colors down to 4) as mask-image for the generator. The area I choose has a lot of trees, so in the end Worldtool generated 1.6 mio single trees (wow).
Posting Permissions
Reply With Quote
