YOLO.py

1. 코드블럭

import cv2
import numpy as np

def draw_text(img, text,
          font=cv2.FONT_HERSHEY_PLAIN,
          pos=(0, 0),
          font_scale=1,
          font_thickness=1,
          text_color=(0, 0, 0),
          text_color_bg=(0, 0, 0)
          ):

    x, y = pos
    text_size, _ = cv2.getTextSize(text, font, font_scale, font_thickness)
    text_w, text_h = text_size
    cv2.rectangle(img, pos, (x + text_w, y + text_h), text_color_bg, -1)
    cv2.putText(img, text, (x, y + text_h + font_scale - 1), font, font_scale, text_color, font_thickness)

    return text_size

min_confidence = 0.5

# Load Yolo
net = cv2.dnn.readNet("machineVision/YOLO/yolov3.weights", "machineVision/YOLO/yolov3.cfg")
classes = []
with open("machineVision/YOLO/coco.names", "r") as f:
    classes = [line.strip() for line in f.readlines()]
layer_names = net.getLayerNames()
output_layers = [layer_names[i[0] - 1] for i in net.getUnconnectedOutLayers()]
colors = np.random.uniform(0, 255, size=(len(classes), 3))

# Loading image
img = cv2.imread("machineVision/YOLO/yolo_03.jpg")
img = cv2.resize(img, None, fx=0.4, fy=0.4)
height, width, channels = img.shape
cv2.imshow("Original Image", img)

# Detecting objects
blob = cv2.dnn.blobFromImage(img, 0.00392, (416, 416), (0, 0, 0), True, crop=False)

net.setInput(blob)
outs = net.forward(output_layers)

# Showing informations on the screen
class_ids = []
confidences = []
boxes = []

for out in outs:
    for detection in out:
        scores = detection[5:]
        class_id = np.argmax(scores)
        confidence = scores[class_id]
        if confidence > min_confidence:
            # Object detected
            center_x = int(detection[0] * width)
            center_y = int(detection[1] * height)
            w = int(detection[2] * width)
            h = int(detection[3] * height)

            # Rectangle coordinates
            x = int(center_x - w / 2)
            y = int(center_y - h / 2)

            boxes.append([x, y, w, h])
            confidences.append(float(confidence))
            class_ids.append(class_id)

indexes = cv2.dnn.NMSBoxes(boxes, confidences, min_confidence, 0.4)
font = cv2.FONT_HERSHEY_PLAIN
font_scale = 1
font_thickness = 1
for i in range(len(boxes)):
    if i in indexes:
        x, y, w, h = boxes[i]
        label = str(classes[class_ids[i]])
        print(i, label)
        color = colors[i]
        # 사각형 그리기
        cv2.rectangle(img, (x, y), (x + w, y + h), color, 2)


        # 사각형 위 채우기
        # cv2.rectangle(img, (x-1, y-15 ), (x-1 + w+2, y), color, cv2.FILLED)
        # cv2.putText(img, label, (x, y-2), font, 1, (0, 255, 0), 1)

        # draw_text함수 사용하여 코딩(맨위에 있음)
        # https://stackoverflow.com/questions/60674501/how-to-make-black-background-in-cv2-puttext-with-python-opencv
        #draw_text(img, label, font=font, pos=(x-1, y-10),text_color_bg=color)
        

        # draw_text함수에서 일부 가져와서 글자 쓰기
        # text_size, _ = cv2.getTextSize(label, font, font_scale, font_thickness)
        # text_w, text_h = text_size
        # cv2.rectangle(img, (x-1, y-11), (x-1 + text_w, y-11 + text_h), color, -1)
        cv2.putText(img, label, (x, y-2), font, 1, (0, 255, 0), 1)
        

cv2.imshow("YOLO Image", img)
cv2.waitKey(0)
cv2.destroyAllWindows()

2. 원본 화면 - 결과 화면

YOLO.zip

4.21MB

 

※ 참고

 i) Coco dataset https://cocodataset.org/#keypoints-2020

COCO - Common Objects in Context

  - YOLO3 부터 coco dataset을 사용하기 시작했으며, coco에는 이미지를 분류하는 카테고리가 80개 정도 있음

  - 그에 반해, PASCAL VOC dataset은 YOLO1 ~ 2에서 사용된 데이터셋으로 분류 카테고리가 20개임

  - ★ YOLO에 대한 설명은 다른 글을 참고바람