{"id":3951,"date":"2020-08-29T03:25:37","date_gmt":"2020-08-28T18:25:37","guid":{"rendered":"https:\/\/now0930.pe.kr\/wordpress\/?p=3951"},"modified":"2020-08-29T03:26:10","modified_gmt":"2020-08-28T18:26:10","slug":"learning-opencv-4-computer-vision-with-python-3-object-detecion","status":"publish","type":"post","link":"https:\/\/now0930.pe.kr\/wordpress\/learning-opencv-4-computer-vision-with-python-3-object-detecion\/","title":{"rendered":"Learning OpenCV 4 Computer Vision with Python 3, object detecion"},"content":{"rendered":"\n

MeanShift\uc640 CamShift\ub85c \ubb3c\uccb4\ub97c \ucd94\uc801\ud560 \uc218 \uc788\ub2e4. \ubc30\uacbd\uc744 \uc81c\uac70\ud558\uae30 \uc704\ud574 backprojection\uc744 \uc5b4\ub5bb\uac8c \uc0ac\uc6a9\ud558\ub294\uc9c0 \uc54c\uc544\uc57c \ud588\ub2e4. <\/p>\n\n\n\n

back projection\uc740 \ubc30\uacbd \ub4f1 \ubcc0\ud654\uac00 \uc5c6\ub294 \uc774\ubbf8\uc9c0\ub97c \ub9c8\uc2a4\ud0b9\ud558\uae30 \uc704\ud574 \uc0ac\uc6a9\ud55c\ub2e4. \ub2e4\uc74c \ub3d9\uc601\uc0c1\uc744 \ubcf4\uba74 \uc54c\uae30 \uc27d\ub2e4. cv2.imshow\ub294 \uc5ec\ub7ec \ucc3d\uc744 \ubcf4\uc5ec\uc8fc\uc9c0 \ubabb\ud55c\ub2e4. matplot.pyplot\uc744 \uc0ac\uc6a9\ud588\ub2e4. opencv\uc758 \uc0c9\uacfc matplot.pyplot \uc0c9\uc774 \ub2e4\ub974\uae30 \ub54c\ubb38\uc5d0 \ubcc0\uacbd\ud574\uc57c \uc81c\ub300\ub85c \ud45c\uc2dc\ub41c\ub2e4.<\/p>\n\n\n\n

class BackProjection(object):\n    def __init__(self):\n        self._imgPath = \".\/20160910_134843_2.jpg\"\n        self._img = cv2.imread(\".\/20160910_134843_2.jpg\",cv2.IMREAD_COLOR)\n        self._roi = cv2.imread(\".\/roi2.jpg\")\n\n        scale_percent = 40 # percent of original size\n        self._width = int(self._img.shape[1] * scale_percent \/ 100)\n        self._height = int(self._img.shape[0] * scale_percent \/ 100)\n        self._dim = (self._width, self._height) \n\n        self._hsv = cv2.cvtColor(self._img, cv2.COLOR_BGR2HSV)\n        self._roi_hsv = cv2.cvtColor(self._roi, cv2.COLOR_BGR2HSV)\n        self._roi_hist = cv2.calcHist([self._roi_hsv],[0,1], None, [180,256],[0,180,0,256])\n        self._mask = cv2.calcBackProject([self._hsv], [0,1], self._roi_hist, [0, 180, 0, 256],1)\n\n\n        #kernel\n        kernel = cv2.getStructuringElement(cv2.MORPH_ELLIPSE, (5,5))\n        self._mask = cv2.filter2D(self._mask, -1, kernel)\n\n        _, self._mask = cv2.threshold(self._mask, 50, 255, cv2.THRESH_BINARY)\n        self._mask = cv2.merge((self._mask, self._mask, self._mask))\n        self._result = cv2.bitwise_and(self._img, self._mask)\n\n\n    def run(self):\n        cv2.resize(self._img,self._dim,interpolation=cv2.INTER_AREA)\n        plt.figure(1)\n        plt.imshow(self._mask)\n        plt.figure(2)\n        #matplot\uc73c\ub85c \ud45c\uc2dc\ud558\uba74 BRG\uc744 RGB\ub85c \ubc14\uafd4\uc57c \ub428\n        plt.imshow(cv2.cvtColor(self._img, cv2.COLOR_BGR2RGB))\n        plt.figure(3)\n        plt.imshow(cv2.cvtColor(self._result, cv2.COLOR_BGR2RGB))\n        plt.show()\n\nif __name__ == '__main__':\n    BackProjection().run()<\/pre>\n\n\n\n
\"\"
\uc6d0\ubcf8 \uc774\ubbf8\uc9c0<\/figcaption><\/figure>\n\n\n\n
\"\"
masking \uc774\ubbf8\uc9c0<\/figcaption><\/figure>\n\n\n\n
\"\"
roi \ud788\uc2a4\ud2b8\uadf8\ub7a8, hsv<\/figcaption><\/figure>\n\n\n\n
\"\"
\uacb0\uacfc \uc774\ubbf8\uc9c0<\/figcaption><\/figure>\n\n\n\n
\uacb0\uacfc\ub97c \ubcf4\uba74 \ubc30\uacbd\uc744 \uae54\ub054\ud558\uac8c \uc9c0\uc6b0\uc9c0 \ubabb\ud588\ub2e4. \uc774\ubbf8\uc9c0 \uc5b4\ub290 \ubd80\ubd84\uc744 roi_hsv\ub85c \uc124\uc815\ud558\ub0d0\uc5d0 \ubc30\uacbd\uc744 \uc5b4\ub5bb\uac8c \uc9c0\uc6b8\uc9c0 \uacb0\uc815\ud55c\ub2e4. \ub354 \ub098\uac00 camshift\ub85c \ubb3c\uccb4\ub97c \ucd94\uc801\ud560 \uc218 \uc788\ub2e4. \uadf8\ub7ec\ub098 \uac12\uc744 \uc5b4\ub5bb\uac8c \ub123\uace0 \ubc30\uacbd\uc774 \uc5b4\ub5a4\uc9c0\uc5d0 \ub530\ub77c \ub9ce\uc740 \uc601\ud5a5\uc744 \ubc1b\ub294\ub2e4. \uacb0\uad6d\uc740 \ub178\uac00\ub2e4..<\/p>\n\n\n\n
\n