Tire Identification Number Detection and Recognition

Requirements

Tested on Ubuntu 18.04.

Direct dependencies:
TCLAP : libtclap-dev (official repo)
Opencv 4.1.2

OpenCV was built from sources (https://github.com/opencv/opencv/tree/4.1.2 and contribs https://github.com/opencv/opencv_contrib/tree/4.1.2).
It must have Tesseract enabled
Tesseract : libtesseract4 libtesseract-dev (official repo)

HOW TO

Sample for detection and recognition :

cd [LocalPath]/samples/
mkdir build_DetectionRecognition
cd build_DetectionRecognition
cmake ../DetectionRecognition -DCMAKE_BUILD_TYPE=Release -DOPENCV_VERSION=4 -DTIN_DR_WITH_VERBOSE=ON
make -j12

Execution :

./DetectionRecognition -i ../../data/tin_001.jpeg -m ../../data/frozen_east_text_detection.pb -c 0.1 -n 0.1 -s -p 10

Help :

-o <string>,  --outputPath <string>
    Path to a directory to save detected text patches

-p <int>,  --padding <int>
    Add padding to detected area

-s,  --useSlidingWindow
    Instead of resizing the input image to 320x320, use a sliding window
    and merge detections

-n <float>,  --nmsThreshold <float>
    Non-maximum suppression threshold

-c <float>,  --confidenceThreshold <float>
    Confidence threshold

-m <string>,  --eastModelPath <string>
    (required)  Path to a binary .pb containing the trained network from
    :

    https://github.com/argman/EAST => EAST: An Efficient and Accurate
    Scene Text Detector (https://arxiv.org/abs/1704.03155v2)

-i <string>,  --inputPath <string>
    (required)  Path to an image