[AutoIt] Auto Fish Bot - Black Desert

[calipso888]

Wandalizatorka88, bglee, they might know a few thigs about xingncode, The manage to baypase it for Cabal Online, I presume they could do it for this one too, but as far I know they have no interest in this game. ��

[killzone]

There's no need to bypass Xigncode for this to work.
What's currently stopping the script isnt Xigncode.
Its the Captcha.

[s3anlike]

This captcha to get it work will kill me X,x

[ahmedwork]

just wanted to share my latest efforts to get a clean captcha window without any noise letters are clear as u can see :3 captcha solving using c# is almost done

[killzone]

Yea, OCR can do that pretty much.
I made a similar application before when I was playing BDO KR.
To help me translate KR texts. xD

[iNiperx]

Originally Posted by ahmedwork

just wanted to share my latest efforts to get a clean captcha window without any noise letters are clear as u can see :3 captcha solving using c# is almost done

Since I dropped the project long time ago, I don't exactly remember what steps I did on photoshop so I can replicate it in C#

This is as close as I've been able to get:


I feel like there's a simpler way and I'm just over doing it.. and it doesn't look good enough x.x

Edit:
Got it in less steps now.. now I have to put it in C# xD

[ahmedwork]

Originally Posted by iNiperx

Since I dropped the project long time ago, I don't exactly remember what steps I did on photoshop so I can replicate it in C#

Edit:
Got it in less steps now.. now I have to put it in C# xD

don't waste your time solving the captcha with ocr unfortunately success rate will be too low because of the background noise even with huge filtering with EmguCV and other tools some samples will always fail :/

[s3anlike]

is there a chance to solve it with autoit yet?

[h96vy]

Originally Posted by ahmedwork

just wanted to share my latest efforts to get a clean captcha window without any noise letters are clear as u can see :3 captcha solving using c# is almost done

i did this too, working real time. not taking a pecture!
we actually walking together not working together XD
I'm so close now

[iNiperx]

Originally Posted by ahmedwork

don't waste your time solving the captcha with ocr unfortunately success rate will be too low because of the background noise even with huge filtering with EmguCV and other tools some samples will always fail :/

Cant you use ImageSearch to find those white arrows on your application?

[killzone]

Originally Posted by iNiperx

Cant you use ImageSearch to find those white arrows on your application?

I think the way he did it is:
Put a form overlay to X/Y (location of Captcha).
Search for Pixel colors that the captcha uses (or simply, If not White = MySearchedColor)
Then, print the found / Matching color.
To make sure and remove the captcha noises, he can contrast it to make it black or white.
After that, he can now read clearly what the colors are.
Maybe at this point he can now use the Tesseract / OCR to distinguish the "Letters".
Then, he'll just put it this way:
If Textbox.text = S then
Send Key "S"
ElseIf Textbox.text = W then
Send Key "W"
And so on...

There's a pain in the *** way as well.
Its longer and needs a lot of images.
What you will do is take a screenshot of EACH CAPTCHA and print screen it.
Then dissect each Letter by Colors.
For example:
Purple S, Red S, White S, Green S and so on...
Then on your screen you search image for those "S"'s and if Found, press S.

Again, that is not an optimal way but it should also work as intended.
It may be longer that way.

[ahmedwork]

Originally Posted by killzone

I think the way he did it is:
Put a form overlay to X/Y (location of Captcha).
Search for Pixel colors that the captcha uses (or simply, If not White = MySearchedColor)
Then, print the found / Matching color.
To make sure and remove the captcha noises, he can contrast it to make it black or white.
After that, he can now read clearly what the colors are.
Maybe at this point he can now use the Tesseract / OCR to distinguish the "Letters".
Then, he'll just put it this way:
If Textbox.text = S then
Send Key "S"
ElseIf Textbox.text = W then
Send Key "W"
And so on...
It may be longer that way.

i've used EmguCv with some option to get this result but unfortunately you won't get the same result with all captcha colors that's why i'm trying a different methods now to get the optimal and fastest way to recognize it
btw i'm currently thinking about count a specified color at 6*6 pixels and if the result is for example 50 then "W" 60 then A , etc

[killzone]

Hmm, then it would be a lot easier if you just search the image for: ▲▼►◄
Or better, If anyone has good knowledge of Reversing, just crack the R23 Bot. lmao

[ragman12]

I am not at your guys level but in actool I could test a pixel. I noticed that the very tip of the triangle up, right, left, and down, is in a different location and is unique to each direction.

if pixel point Up = red green yellow blue white or grey then [up],

if pixel right = red green yellow blue white or grey then [right]

if pixel down = red green yellow blue white or grey then [down]

if pixel left = red green yellow blue white or grey then [left]

I know it is not that simple. But I do think that you don't have to grab the whole picture. if a specific pixle has a qualifying color then the bot can act. that is all I am saying; perhaps not. just a thought.

If you do get something working I would like to know in my inblox thnkx .

also isnt there someway to create a mob killer that searches for that bar that every mob has above its head? or could drive him self towards red dots on the radar and aoes when he get close ?

[jacktyme]

I got the space and perfect files to work. The autoperfect always gets the space bar part right.

As like everyone else I am stuck on the captcha part.

I think this part is all we have to change, we don't have to change anything in the coords.txt file:

If ($xl == 0) Then
$Res = _ImageSearchArea("checker.bmp", 0, 630, 412, 713, 467, $xl, $yl, 60)
$Res = _ImageSearchArea("checker3.bmp", 0, 630, 412, 713, 467, $xl, $yl, 60)
$Res = _ImageSearchArea("checker2.bmp", 0, 630, 412, 713, 467, $xl, $yl, 60)
$Res = _ImageSearchArea("checker4.bmp", 0, 630, 412, 713, 467, $xl, $yl, 60)
If ($xl > 0)  Then
Sleep(800)
$Res = _ImageSearchArea("checker.bmp", 0, 630, 412, 713, 467, $xl, $yl, 60)
$Res = _ImageSearchArea("checker4.bmp", 0, 630, 412, 713, 467, $xl, $yl, 60)
$Res = _ImageSearchArea("checker3.bmp", 0, 630, 412, 713, 467, $xl, $yl, 60)
$Res = _ImageSearchArea("checker2.bmp", 0, 630, 412, 713, 467, $xl, $yl, 60)

The 630 after the 0 is I think the first pixel coord and the 412 is the Y coord.
The code is searching for the checker.bmp image. Now one reason I think it doesn't work is the image of the bar isn't the same as the bar in game. Maybe that is the way the russian version bar looks like. This is just a theory.


The 713, 467 could be the default desktop resolution or game resolution.

Someone in the thread has gotten it to actually type characters.

It would be great to get this going.

Any help appreciated.