OpenCV Python 고급 활용 강좌 소스 코드 4

 

적응형 이진화

 

import cv2

src = cv2.imread("tree.jpg")
gray = cv2.cvtColor(src, cv2.COLOR_BGR2GRAY)
binary = cv2.adaptiveThreshold(gray, 255, cv2.ADAPTIVE_THRESH_MEAN_C, cv2.THRESH_BINARY, 467, 37)

cv2.imshow("binary", binary)
cv2.waitKey(0)
cv2.destroyAllWindows()

 

 

템플릿 매칭

 

import cv2

src = cv2.imread("hats.png", cv2.IMREAD_GRAYSCALE)
templit = cv2.imread("hat.png", cv2.IMREAD_GRAYSCALE)
dst = cv2.imread("hats.png")

result = cv2.matchTemplate(src, templit, cv2.TM_SQDIFF_NORMED)

minVal, maxVal, minLoc, maxLoc = cv2.minMaxLoc(result)
x, y = minLoc
h, w = templit.shape

dst = cv2.rectangle(dst, (x, y), (x +  w, y + h) , (0, 0, 255), 1)
cv2.imshow("dst", dst)
cv2.waitKey(0)
cv2.destroyAllWindows()

 

 

ORB(Oriented FAST and Rotated BRIEF

 

import cv2
import numpy as np

src = cv2.imread("apple_books.jpg")
gray = cv2.cvtColor(src, cv2.COLOR_BGR2GRAY)
target = cv2.imread("apple.jpg", cv2.IMREAD_GRAYSCALE)

orb = cv2.ORB_create(
    nfeatures=40000,
    scaleFactor=1.2,
    nlevels=8,
    edgeThreshold=31,
    firstLevel=0,
    WTA_K=2,
    scoreType=cv2.ORB_HARRIS_SCORE,
    patchSize=31,
    fastThreshold=20,
)

kp1, des1 = orb.detectAndCompute(gray, None)
kp2, des2 = orb.detectAndCompute(target, None)

bf = cv2.BFMatcher(cv2.NORM_HAMMING, crossCheck=True)
matches = bf.match(des1, des2)
matches = sorted(matches, key=lambda x: x.distance)

for i in matches[:100]:
    idx = i.queryIdx
    x1, y1 = kp1[idx].pt
    cv2.circle(src, (int(x1), int(y1)), 3, (255, 0, 0), 3)

cv2.imshow("src", src)
cv2.waitKey()

 

 

마우스 콜백

 

import cv2
import numpy as np

def mouse_event(event, x, y, flags, param):
    global radius
    
    if event == cv2.EVENT_FLAG_LBUTTON:    
        cv2.circle(param, (x, y), radius, (255, 0, 0), 2)
        cv2.imshow("draw", src)

    elif event == cv2.EVENT_MOUSEWHEEL:
        if flags > 0:
            radius += 1
        elif radius > 1:
            radius -= 1

radius = 3
src = np.full((500, 500, 3), 255, dtype=np.uint8)

cv2.imshow("draw", src)
cv2.setMouseCallback("draw", mouse_event, src)
cv2.waitKey()

 

리매핑

 

import cv2
import numpy as np

src = cv2.imread("buildings.jpg")
height, width = src.shape[:2]
map2, map1 = np.indices((height, width), dtype=np.float32)

map1 = map1 + width / 100 * np.sin(map1)
map2 = map2 + height / 100 * np.cos(map2)

dst = cv2.remap(src, map1, map2, cv2.INTER_CUBIC)
cv2.imshow("dst", dst)
cv2.waitKey()

 

색상 맵

 

import cv2

src = cv2.imread("beach.jpg")
dst = cv2.applyColorMap(src, cv2.COLORMAP_OCEAN)

cv2.imshow("dst", dst)
cv2.waitKey()
cv2.destroyAllWindows()

 

K-평균 군집화 알고리즘

 

import numpy as np
import cv2

src = cv2.imread("flower.jpg")

data = src.reshape(-1, 3).astype(np.float32)
criteria = (cv2.TERM_CRITERIA_MAX_ITER + cv2.TERM_CRITERIA_EPS, 10, 0.001)
retval, bestLabels, centers = cv2.kmeans(data, 5, None, criteria, 10, cv2.KMEANS_RANDOM_CENTERS)

centers = centers.astype(np.uint8)
dst = centers[bestLabels].reshape(src.shape)

cv2.imshow("dst", dst)
cv2.waitKey()
cv2.destroyAllWindows()

 

 

K-평균 최근접 알고리즘

 

import cv2
import numpy as np


def loadTrainData(image_path, label_path):
    with open(image_path, "rb") as image_data:
        images = np.frombuffer(image_data.read(), dtype=np.uint8, offset=16)
    with open(label_path, "rb") as label_data:
        labels = np.frombuffer(label_data.read(), dtype=np.uint8, offset=8)
    return images.reshape(-1, 784), labels


train_x, train_y = loadTrainData(
    "./fashion-mnist/train-images-idx3-ubyte",
    "./fashion-mnist/train-labels-idx1-ubyte"
)
test_x, test_y = loadTrainData(
    "./fashion-mnist/t10k-images-idx3-ubyte",
    "./fashion-mnist/t10k-labels-idx1-ubyte"
)

label_dict = {
    0: "T-shirt/top",
    1: "Trouser",
    2: "Pullover",
    3: "Dress",
    4: "Coat",
    5: "Sandal",
    6: "Shirt",
    7: "Sneaker",
    8: "Bag",
    9: "Ankle boot",
}

knn = cv2.ml.KNearest_create()
retval = knn.train(train_x.astype(np.float32), cv2.ml.ROW_SAMPLE, train_y.astype(np.int32))

count = 500
retval, results, neighborResponses, dist = knn.findNearest(
    test_x[:count].astype(np.float32), k=7
)

matches = results.astype(np.uint8) == test_y[:count][:, None]
print(np.count_nonzero(matches) / count * 100)

for idx, result in enumerate(results):
    print("Index : {}".format(idx))
    print("예측값 : {}".format(label_dict[int(result)]))
    print("실제값 : {}".format(label_dict[test_y[idx]]))
    cv2.imshow("images", test_x[idx].reshape(28, 28, 1))
    cv2.waitKey()