FRLG Summary stat reading initial implementation#1130
Open
miniumknight wants to merge 12 commits intoPokemonAutomation:mainfrom
Open
FRLG Summary stat reading initial implementation#1130miniumknight wants to merge 12 commits intoPokemonAutomation:mainfrom
miniumknight wants to merge 12 commits intoPokemonAutomation:mainfrom
Conversation
pifopi
reviewed
Mar 11, 2026
SerialPrograms/Source/PokemonFRLG/Programs/TestPrograms/PokemonFRLG_ReadStats.cpp
Outdated
Show resolved
Hide resolved
SerialPrograms/Source/PokemonFRLG/Inference/PokemonFRLG_StatsReader.h
Outdated
Show resolved
Hide resolved
SerialPrograms/Source/PokemonFRLG/Inference/PokemonFRLG_DigitReader.cpp
Outdated
Show resolved
Hide resolved
SerialPrograms/Source/PokemonFRLG/Inference/PokemonFRLG_DigitReader.cpp
Outdated
Show resolved
Hide resolved
SerialPrograms/Source/PokemonFRLG/Inference/PokemonFRLG_DigitReader.cpp
Outdated
Show resolved
Hide resolved
SerialPrograms/Source/PokemonFRLG/Inference/PokemonFRLG_DigitReader.cpp
Outdated
Show resolved
Hide resolved
SerialPrograms/Source/PokemonFRLG/Inference/PokemonFRLG_DigitReader.h
Outdated
Show resolved
Hide resolved
SerialPrograms/Source/PokemonFRLG/Inference/PokemonFRLG_StatsReader.cpp
Outdated
Show resolved
Hide resolved
Mysticial
reviewed
Mar 12, 2026
| // | ||
| // Returns the parsed integer, or -1 on failure. | ||
| int read_digits_waterfill_template( | ||
| Logger& logger, |
Mysticial
reviewed
Mar 12, 2026
|
|
||
| // Read a string of decimal digits from `stat_region` by splitting the region into | ||
| // a fixed number of equal-width segments, instead of using waterfill. | ||
| // Useful when digits are tightly packed or overlapping and waterfill merges them. |
Collaborator
There was a problem hiding this comment.
When are the digit overlapping?
Mysticial
reviewed
Mar 12, 2026
| std::string prefix = "DebugDumps/ocr_" + label + "_" + std::to_string(id); | ||
|
|
||
| // Save raw input | ||
| image.save(prefix + "_0_raw.png"); |
Collaborator
There was a problem hiding this comment.
Put this behind some sort of debug flag.
Mysticial
reviewed
Mar 12, 2026
| cv::Mat src = image.to_opencv_Mat(); | ||
|
|
||
| // Step 1: Gaussian blur at NATIVE resolution with 5×5 kernel. | ||
| // The 5×5 kernel reaches 2 pixels away (vs 1px for 3×3), bridging |
Collaborator
There was a problem hiding this comment.
Let's try to avoid non-ascii characters in source code.
Mysticial
reviewed
Mar 12, 2026
| // Step 3: BW threshold on the smooth upscaled image. | ||
| ImageRGB32 bw = | ||
| to_blackwhite_rgb32_range(resized_img, in_range_black, bw_min, bw_max); | ||
| bw.save(prefix + "_3_bw.png"); |
Collaborator
There was a problem hiding this comment.
Same as earlier, put all the temporary/debug images behind a debug flag.
Mysticial
reviewed
Mar 12, 2026
| // | ||
| // The native blur connects gaps. Post-BW padding provides margins. | ||
| static ImageRGB32 preprocess_for_ocr(const ImageViewRGB32 &image, | ||
| const std::string &label, |
Collaborator
There was a problem hiding this comment.
Nit: If a function headers needs to be broken up, we move all of it on the following lines such that they start with one indent rather than mid-way through the page.
ReturnType function_name(
Type param0, Type param1,
Type param2, Type param3,
Type param2, Type param3
){
}
Mysticial
reviewed
Mar 12, 2026
| // Read Name (white text with shadow) | ||
| auto name_result = Pokemon::PokemonNameReader::instance().read_substring( | ||
| logger, language, extract_box_reference(game_screen, m_box_name), | ||
| {{combine_rgb(235, 235, 235), combine_rgb(255, 255, 255)}}); |
Collaborator
There was a problem hiding this comment.
This filter range 235 - 255 is probably too small. I've seen whites be as low as 0xc0 (192).
How low can this go before it starts pulling in other stuff we don't want?
If the margin is too small, we typically resort to "multi-filtering" where we run it across a range of narrower bands and pick the best result from all of them.
Mysticial
reviewed
Mar 12, 2026
| if (px.red() > 200 && px.green() > 200 && px.blue() > 200) { | ||
| preprocessed.pixel(c, r) = 0xff000000u; // white text → black | ||
| } else { | ||
| preprocessed.pixel(c, r) = level_box.pixel(c, r); // keep as-is |
Collaborator
There was a problem hiding this comment.
Can this entire double-loop be done using this call?
ImageRGB32 preprocessed = filter_rgb32_range(level_box, 0xffc8c8c8, 0xffffffff, 0xff000000, true);
Mysticial
reviewed
Mar 12, 2026
Collaborator
|
This looks pretty good! Thank you for doing this! |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Basic implementation of FRLG stat reading using either PaddleOCR if enabled or Template matching to set images (From Packages)