main.py

import glob
import os
import sys
import time

import cv2 as cv
from matplotlib import pyplot as plt

from alpd.segmentation.segment import newOldSegmentation
from integeration.client import endPoint
from recognition.KNN import classify_unlabelled_directory
from recognition.KNN import predictChars

path = './outputs/'
# url = "http://192.168.20.51:8080/video"
url = "http://192.168.20.51:8080/video"
# url = "http://9.246.91.33:8080/video"
sys.path.insert(0, './localisation')
import detect as detect
from localisation.core.functions import load_model


def printChars(chars, count):
    if not os.path.isdir(f'{path}{count}'):
        os.makedirs(f'{path}{count}')

    for i in range(len(chars)):
        plt.subplot(1, 10, i + 1)
        plt.imshow(chars[i], cmap='gray')
        cv.imwrite(f'{path}{count}/{i + 1}.png', chars[i])
        plt.axis('off')
    # plt.show()


# Main  which responsible for integration of whole image processing pipelines
if __name__ == '__main__':
    saveModel = load_model()
    while True:

        try:
            countPlate = 1
            # stream(url)

            # localization
            start = time.time()
            detect.crop_multiple("./localisation/data/demo/", False, saveModel)
            # detect_and_crop_from_video("./localisation/data/videos/1.mp4", saveModel)
            for filename in os.scandir("./detections/"):
                try:
                    # segmentation
                    # _, chars = segmentChars(filename.path)
                    chars = newOldSegmentation(filename.path)
                    # chars = segmentCharacters(filename.path)
                    # print chars segmented
                    printChars(chars, countPlate)

                    # recognition
                    predicted_chars = classify_unlabelled_directory(f'{path}{countPlate}/')

                    end = time.time()
                    # print string
                    lp = predictChars(predicted_chars)
                    print(lp)
                    print(end - start)
                    countPlate += 1

                    # send string to middle-ware
                    print(endPoint(lp))
                    files = glob.glob('./detections/*')
                    for file in files:
                        os.remove(file)
                    files = glob.glob('./outputs/1/*')
                    for file in files:
                        os.remove(file)
                    files=glob.glob('./green_boxes/*')
                    for file in files:
                        os.remove(file)

                except Exception:
                    # print("error")
                    continue

        except Exception:
            # print("outer error")
            continue

# segObject = Segmentation(filename.path)
# chars = segObject.run()
# segObject = Segmentation(filename.path)
# len1, chars1 = segmentChars(filename.path)
# len2, chars2 = segObject.run()
# mx = max(len1, len2)
# chars = chars1
# if mx == len1 and len1 <= 7:
#     print("================== First ===================")
# else:
#     print("================== Second ===================")
#     chars = chars2