本文共 5191 字，大约阅读时间需要 17 分钟。
#! /usr/bin/env pythonfrom cProfile import Profileimport pstatsimport numpy as npimport cv2import randomimport timeimport matplotlib.pyplot as pltdefaultNbSamples = 20defaultReqMatches = 2defaultRadius = 20defaultSubsamplingFactor = 16background = 0foreground = 255def Initial_ViBe(gray, samples): # gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY) height = gray.shape[0] width = gray.shape[1] print("*****Initlize Time******") start1 = time.clock() for i in range(height): for j in range(width): for k in range(defaultNbSamples): rand = round(random.uniform(-1, 1)) r = i + rand if r < 0: r = 0 if r > height: r = height - 1 rand = round(random.uniform(-1, 1)) c = j + rand if c < 0: c = 0 if c > width: c = width - 1 samples[i][j][k] = gray[i][j] samples[i][j][defaultNbSamples - 1] = 0 end = time.clock() print(end - start1) print("******************************") return samplesdef update(gray, samples): height = gray.shape[0] width = gray.shape[1] # print(height, width) # segMat = np.zeros((gray.shape[0], gray.shape[1], 1), np.uint8) segMat = np.zeros((gray.shape[0], gray.shape[1], 1)) for i in range(height): for j in range(width): count = 0 idex = 0 dist = 0 # 遍历每个像素，判断与背景样本的相似程度 while count < defaultReqMatches and idex < defaultNbSamples: dist = abs(gray[i][j] - samples[i][j][idex]) if dist < defaultRadius: # 统计相似度小于阈值的个数 count = count + 1 idex = idex + 1 # 大于#min，则认为该点为背景点 if count >= defaultReqMatches: # 判断为背景像素，只有背景点才可以被用来传播和更新存储样本值 samples[i][j][defaultNbSamples - 1] = 0 # 输出二值图像用 segMat[i][j] = background rand = round(random.uniform(0, defaultSubsamplingFactor - 1)) if rand == 0: rand = round(random.uniform(0, defaultNbSamples - 1)) samples[i][j][rand] = gray[i][j] rand = round(random.uniform(0, defaultSubsamplingFactor - 1)) if rand == 0: rand = round(random.uniform(-1, 1)) iN = i + rand # 以下是防止越界 if iN < 0: iN = 0 if iN > height: iN = height - 1 rand = round(random.uniform(-1, 1)) jN = j + rand if jN < 0: jN = 0 if jN > width: jN = width - 1 rand = round(random.uniform(0, defaultNbSamples - 1)) samples[i][j][rand] = gray[i][j] else: segMat[i][j] = foreground samples[i][j][defaultNbSamples - 1] = samples[i][j][defaultNbSamples - 1] + 1 if samples[i][j][defaultNbSamples - 1] > 50: rand = round(random.uniform(0, defaultNbSamples - 1)) if rand == 0: rand = round(random.uniform(0, defaultNbSamples - 1)) samples[i][j][rand] = gray[i][j] return segMat, samplesdef main(): vc = cv2.VideoCapture("video.avi") c = 0 if vc.isOpened(): rval, frame = vc.read() else: rval = False samples = np.zeros((500, 500, defaultNbSamples)) while rval: rval, frame = vc.read() if c == 1: # 将输入转为灰度图 gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY) samples = Initial_ViBe(gray, samples) else: # 将输入转为灰度图 gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY) # 输出二值图 (segMat, samples) = update(gray, samples) # 转为uint8类型 segMat = np.array(segMat, np.uint8) # 形态学处理模板初始化 kernel1 = cv2.getStructuringElement(cv2.MORPH_RECT, (7, 7)) # 开运算 opening = cv2.morphologyEx(segMat, cv2.MORPH_OPEN, kernel1) # 形态学处理模板初始化 kernel2 = cv2.getStructuringElement(cv2.MORPH_RECT, (7, 7)) # 闭运算 closed = cv2.morphologyEx(segMat, cv2.MORPH_CLOSE, kernel2) # 寻找轮廓 Img, contours, hierarchy = cv2.findContours( closed, cv2.RETR_TREE, cv2.CHAIN_APPROX_SIMPLE) for i in range(0, len(contours)): x, y, w, h = cv2.boundingRect(contours[i]) print(w * h) if w * h > 400 and w * h < 10000: cv2.rectangle(frame, (x, y), (x + w, y + h), (0, 255, 0), 2) cv2.imshow("frame", frame) cv2.imwrite("./result/" + str(c) + ".jpg", frame, [int(cv2.IMWRITE_PNG_STRATEGY)]) k = cv2.waitKey(1) if k == 27: vc.release() cv2.destroyAllWindows() c = c + 1main()
转载地址：http://syhws.baihongyu.com/