Thanks for the answer.
Still, didn't get one part.
You are saying that on that last tiny part it will always lose the lock? why is that?
For example on fatdoobie version we have 3 bmp with red bar at 25, 50, full, 75...
The problem would be fixed by adding the 5-10% health bar?
Or i am missing something?
(i know this exactly one doesn't work with the 3 red bar bmp)
Well... this version should be nearly twice as fast as the other script because this one only does 1 imagesearch where as the other script does two image searches.
This script works by searching for the following image:
BUT the white space is ignored and treated as transparent space so it only looks for the red area & the grey areas on the center part of the screen (for speed).
If a monsters life is lower than the amount of red pixels shown in this image (9px) then it won't be finding this image on your screen and as such the lock on will vanish.
If you adjusted the image so the red section was only 4px~ wide then it increases its sensitivity and should find the health bar more (when it's lower), but it also increases the chances of more false positives.
You'd have to adjust the offset by 5px to make sure you are aiming at the correct location.
It's looking for this part of your screen:
Bare in mind look at the differences in the colors based on certain colored backgrounds, also note these are at different distances so it makes it much harder to find 1 solution to cover all scenarios.
The exact part it's looking for is:
Which as you can see when it's overlayed on the screenshots that I have taken, it doesn't match perfectly for all 3 and it certainly can be improved on a 1080p display.
See my attached set of images, try out all of these images... I suspect the Test4 and Test5 ones will be the best but I might be wrong, you'd have to adjust your offset because these pictures are different dimensions. I think the adjustments will be an increase in +7px for the x axis and a +2px increase for the y axis.
Issues I can see happening:
- Same 5% issue (Slightly smaller chance of this happening...)
- Same issue where some monsters aren't detected because of the background
Let me know how it goes
Last edited by RandomGuyRandomTime; 10-01-2018 at 11:19 AM.
●▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬●
▂▃▅▇█ RandomGuyRandomTime 's Vouches █▇▅▃▂
●▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬●
abolins6 (10-05-2018),Bf109Ace (10-24-2018),bighape (10-02-2018),Forsaga (10-02-2018),ImNotBlacc (10-01-2018),iPSxFEARx3 (10-01-2018),Masbka (10-02-2018),OldTimeGmer (10-02-2018),ricardmark1 (10-01-2018),Shnking90 (10-02-2018),siyu9469 (10-02-2018),UnrealityPrime (10-02-2018),yashspartan1 (10-01-2018)
I've looked into autoit before but there were two things that turned me away. Unless things have changed, I do not believe they have any native image search function. I also believe that variables and functions are a little more dificult to work with. Since ahk is essentially a modified fork, the code is similar but autoit is simpler (if that makes sense). Also, the autoit community is garbage and will only provide assistance if you can prove it's not for game automation and their documentation is pretty lackluster.
- - - Updated - - -
Spot on write up! That is exactly what I'm going for and you nailed the reasoning for false positives and the current issue with extremely low health guardians. I still think fatdobie is onto something with the multi image search but his code searches for 4 images at a time.
Last edited by Ahlwong; 10-02-2018 at 04:33 PM.
●▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬●
▂▃▅▇█ RandomGuyRandomTime 's Vouches █▇▅▃▂
●▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬●
Look forward to testing the images, not sure how long we will all have to wait though :P
Check my last edit, also I just remembered my images won't be working for shielded units. I'll update that later and also try and introduce some images (multiple checks) for yellow enemies too
There's also a series of various shapes which is annoying, I need to do some testing to see whats what.
I'm currently going through places and getting screenshots of random mobs to work out a nice set of images still
- - - Updated - - -
1080p, bright 4. Everything on highest quality.
Last edited by RandomGuyRandomTime; 10-01-2018 at 01:28 PM.
●▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬●
▂▃▅▇█ RandomGuyRandomTime 's Vouches █▇▅▃▂
●▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬●
JackAim (10-01-2018)
Does ingame settings like chromatic aberation and all that affet the bot do you think, because you said you have everything on highest.. not sure what that means exactly when looking at all options in game, also is your render resolution in game at 100% for 1080p or did u higher it for better visuals etc. Just a thought
It's way to slow.
Do i have to rename the picture or just paste it on andrew's?
Also Andrew i find some strange thing, when i try to pull mouse to left from an enemy its impossible(strong lock), but when i pull mouse to right side its loses lock, can it be solved some how?
This is exactly what I was about to post about, I mentioned other day as well, if you strafe it stays locked, if you strafe and move mouse left as you are strafe tracking it stays pushed onto target, if you do the same and strafe and push mouse right of any amount of movement it breaks the whole locking.
what the heck is causing that. This means when in heat of battle you have to always mentally aim at the body of the person you are shooting at as you can not rely on the tracking moving the mouse alone as in game play you are still trying to avoid things at same time as aim bot shooting. its a weird thing.
Last edited by moosenoodles; 10-01-2018 at 02:12 PM.
Do i need to rename the files or just put the pic i want on andrew's folder?
BTW, i am using 800 dpi and 5 sens in game.
It seems so slow the tracking and doesn't stay on target with this one > Test5 - Averages Alt Final.bmp.
Using same settings.ini as you
Any idea why?
Last edited by Ahlwong; 10-02-2018 at 04:33 PM.