egor loh

import helpers
import cv2
import random
import numpy as np

ACC_CONST = 2400
def process(img):
    ans = cv2.resize(img, (64, 64))
    cv2.blur(ans, (2, 2), ans)
    ans = cv2.inRange(ans, (0, 5, 0), (228, 111, 255))
    ans = cv2.morphologyEx(ans, cv2.MORPH_OPEN, (2, 2))
    return ans

def get_standart_signs():
    """Функция, позволяющая получить стандартные изображения знаков для сравнения
    с оригинальными изображениями. Стандартные изображения знаков хранятся во внутреннем каталоге."""
    # стандартные изображения дорожных знаков

    a_unevenness = cv2.imread("data/standards/a_unevenness.jpg")

    no_drive = cv2.imread("data/standards/no_drive.png")

    no_entry = cv2.imread("data/standards/no_entry.jpg")

    parking = cv2.imread("data/standards/parking.jpg")

    pedistrain = cv2.imread("data/standards/pedistrain.jpg")

    road_works = cv2.imread("data/standards/road_works.jpg")

    stop = cv2.imread("data/standards/stop.jpg")

    way_out = cv2.imread("data/standards/way_out.jpg")

    standart_signs = [
        pedistrain,
        no_drive,
        stop,
        way_out,
        no_entry,
        road_works,
        parking,
        a_unevenness
    ]

    for i in range(8):
        standart_signs[i] = process(standart_signs[i])
    return standart_signs

def one_hot_encode(label):

    one_hot_encoded = []
    if label == "none":
        one_hot_encoded = [0, 0, 0, 0, 0, 0, 0, 0]
    elif label == "pedistrain":
        one_hot_encoded = [1, 0, 0, 0, 0, 0, 0, 0]
    elif label == "no_drive":
        one_hot_encoded = [0, 1, 0, 0, 0, 0, 0, 0]
    elif label == "stop":
        one_hot_encoded = [0, 0, 1, 0, 0, 0, 0, 0]
    elif label == "way_out":
        one_hot_encoded = [0, 0, 0, 1, 0, 0, 0, 0]
    elif label == "no_entry":
        one_hot_encoded = [0, 0, 0, 0, 1, 0, 0, 0]
    elif label == "road_works":
        one_hot_encoded = [0, 0, 0, 0, 0, 1, 0, 0]
    elif label == "parking":
        one_hot_encoded = [0, 0, 0, 0, 0, 0, 1, 0]
    elif label == "a_unevenness":
        one_hot_encoded = [0, 0, 0, 0, 0, 0, 0, 1]

    return one_hot_encoded

def predict_label(image):

    standart_signs = get_standart_signs()
    predicted_label = [0, 0, 0, 0, 0, 0, 0, 0]
    image = process(image)

    results = [0 for i in range(len(standart_signs))]

    for i in range(64):
        for j in range(64):
            for k in range(8):
                if image[i][j] == standart_signs[k][i][j]:
                    results[k] += 1
    _max = max(results)
    if _max > ACC_CONST:
        predicted_label[results.index(_max)] = 1
    return predicted_label