UpAllNate / SpokenScreen

Detect on-screen features with pixel analysis. Supports simple scripting in TOML formats for executing multiple profiles.

SpokenScreen

Core pattern detection / audio playback / hash calculation / image collection code

testColors:

White: 255, 255, 255 White to Red (one pixel): 246, 144, 145 Red: 237, 28, 36 Red to Green (one pixel): 135, 103, 56 Green: 34, 177, 76 Green to Orange (one pixel): 145, 152, 58 Orange: 255, 127, 39 Orange to Green (one pixel): 145, 152, 58 Green: 34, 177, 76 Green to Blue (one pixel): 17, 169, 154 Blue: 0, 162, 232 Blue to White (one pixel): 127, 208, 243 White: 255, 255, 255

TODO: Implement Implement loading of screenshot to 2D pixel array / numpy pixel list: https://stackoverflow.com/questions/60293637/read-image-pixels-row-by-row

Pokemon FireRed Operations

Text box detection

1. Take screenshot
2. pixelSequenceScan the center column of pixels, bottom half of the screen for text box colors
3. If not successful, return to screenshot
4. If successful, get pixelSequenceScan of row between column edges ...

stateDiagram-v2
    [*] --> screenShot
    screenShot --> pScan_V
    pScan_V --> screenShot
    pScan_V --> getMidSE_V
    getMidSE_V --> pScan_H
    pScan_H --> screenShot
    pScan_H --> getMidSE_H
    getMidSE_H --> ComputeTBCorners
    ComputeTBCorners --> GetLines
    GetLines --> StitchLinesToSquare
    StitchLinesToSquare --> ComputeHashes
    ComputeHashes --> CheckHash_1
    CheckHash_1 --> screenShot
    CheckHash_1 --> CheckHash_2
    CheckHash_2 --> screenShot
    CheckHash_2 --> PlayAudio

Loading

chh