egullickson/motovaultpro
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

fix: Build errors and tesseract removal
Some checks failed
Deploy to Staging / Build Images (pull_request) Failing after 4m14s
Details
Deploy to Staging / Deploy to Staging (pull_request) Has been skipped
Details
Deploy to Staging / Verify Staging (pull_request) Has been skipped
Details
Deploy to Staging / Notify Staging Ready (pull_request) Has been skipped
Details
Deploy to Staging / Notify Staging Failure (pull_request) Successful in 8s
Details

Browse Source
This commit is contained in:
Eric Gullickson
2026-02-07 12:12:04 -06:00
parent cf114fad3c
commit b9fe222f12
16 changed files with 35 additions and 238 deletions
Download Patch File Download Diff File Expand all files Collapse all files

12
ocr/app/preprocessors/vin_preprocessor.py
View File

@@ -93,7 +93,7 @@ class VinPreprocessor:
gray = cv_image
steps_applied.append("grayscale")
# Upscale small images for better OCR (Tesseract needs ~300 DPI)
# Upscale small images for better OCR (~300 DPI recommended)
gray = self._ensure_minimum_resolution(gray)
steps_applied.append("resolution_check")
@@ -129,14 +129,14 @@ class VinPreprocessor:
)
# Minimum width in pixels for reliable VIN OCR.
# A 17-char VIN needs ~30px per character for Tesseract accuracy.
# A 17-char VIN needs ~30px per character for reliable OCR accuracy.
MIN_WIDTH_FOR_VIN = 600
def _ensure_minimum_resolution(self, image: np.ndarray) -> np.ndarray:
"""
Upscale image if too small for reliable OCR.
Tesseract works best at ~300 DPI. Mobile photos of VINs may have
OCR works best at ~300 DPI. Mobile photos of VINs may have
the text occupy only a small portion of the frame, resulting in
low effective resolution for the VIN characters.
"""
@@ -160,7 +160,7 @@ class VinPreprocessor:
Colored backgrounds have a low min value (e.g. green sticker:
min(130,230,150) = 130) → inverted to 125 (medium gray).
The inversion ensures Tesseract always receives dark-text-on-
The inversion ensures the OCR engine always receives dark-text-on-
light-background, which is the polarity it expects.
"""
b_channel, g_channel, r_channel = cv2.split(bgr_image)
@@ -168,8 +168,8 @@ class VinPreprocessor:
min_channel = np.minimum(np.minimum(b_channel, g_channel), r_channel)
# Invert so white text (min=255) becomes black (0) and colored
# backgrounds (min~130) become lighter gray (~125). Tesseract
# expects dark text on light background.
# backgrounds (min~130) become lighter gray (~125). OCR engines
# expect dark text on light background.
inverted = cv2.bitwise_not(min_channel)
gray = cv2.cvtColor(bgr_image, cv2.COLOR_BGR2GRAY)