全部代碼鏈接

首先效果圖：

語言：python + OpenCV

簡單的背景減差算法實現移動物體檢測。

學習目標：

1. 理解背景減差算法

2. OpenCV圖像濾波

3. 通過連通區域檢測目標

前景 = 當前幀 - 背景

? ? ? ? 但在某些情況下，我們不能得到靜態幀，因為光可以改變，或者一些物體會被某人移動，或者總是存在運動，等等。在這種情況下，我們保存了一些幀，并試圖找出其中大多數像素是相同的，然后這個像素變為背景的一部分。一般來說，我們如何得到這個背景，以及我們用來選擇更合適的濾波器。

在這個例子中，我們使用MOG（混合高斯）進行背景減差，效果：

可以看出一些噪聲，和陰影，我們使用標準的濾波器去移除他們。

import os import logging import logging.handlers import randomimport numpy as np import skvideo.io import cv2 import matplotlib.pyplot as pltimport utils # without this some strange errors happen cv2.ocl.setUseOpenCL(False) random.seed(123)# ============================================================================ IMAGE_DIR = "./out" VIDEO_SOURCE = "input.mp4" SHAPE = (720, 1280) # HxW # ============================================================================def train_bg_subtractor(inst, cap, num=500):'''BG substractor need process some amount of frames to start giving result'''print ('Training BG Subtractor...')i = 0for frame in cap:inst.apply(frame, None, 0.001)i += 1if i >= num:return capdef main():log = logging.getLogger("main")# creting MOG bg subtractor with 500 frames in cache# and shadow detctionbg_subtractor = cv2.createBackgroundSubtractorMOG2(history=500, detectShadows=True)# Set up image source# You can use also CV2, for some reason it not working for mecap = skvideo.io.vreader(VIDEO_SOURCE)# skipping 500 frames to train bg subtractortrain_bg_subtractor(bg_subtractor, cap, num=500)frame_number = -1for frame in cap:if not frame.any():log.error("Frame capture failed, stopping...")breakframe_number += 1utils.save_frame(frame, "./out/frame_%04d.png" % frame_number)fg_mask = bg_subtractor.apply(frame, None, 0.001)utils.save_frame(frame, "./out/fg_mask_%04d.png" % frame_number)# ============================================================================if __name__ == "__main__":log = utils.init_logging()if not os.path.exists(IMAGE_DIR):log.debug("Creating image directory `%s`...", IMAGE_DIR)os.makedirs(IMAGE_DIR)main()

濾波

對于這個例子，我們用到了下列濾波器：

閾值、膨脹、腐蝕、開運算、閉運算

現在我們將使用它們來消除前景掩碼上的一些噪聲。
首先，我們將使用閉運算來消除區域的空隙，然后開運算以移除1到2個PX點，然后在膨脹處理使目標更大。

def filter_mask(img):kernel = cv2.getStructuringElement(cv2.MORPH_ELLIPSE, (2, 2))# Fill any small holesclosing = cv2.morphologyEx(img, cv2.MORPH_CLOSE, kernel)# Remove noiseopening = cv2.morphologyEx(closing, cv2.MORPH_OPEN, kernel)# Dilate to merge adjacent blobsdilation = cv2.dilate(opening, kernel, iterations=2)# thresholdth = dilation[dilation < 240] = 0return th

前景如下：

通過檢測連通區域進行目標檢測

我們使用cv2.findContours函數實現；

cv2.CV_RETR_EXTERNAL — get only outer contours. cv2.CV_CHAIN_APPROX_TC89_L1 - use Teh-Chin chain approximation algorithm (faster) def get_centroid(x, y, w, h):x1 = int(w / 2)y1 = int(h / 2)cx = x + x1cy = y + y1return (cx, cy)def detect_vehicles(fg_mask, min_contour_width=35, min_contour_height=35):matches = []# finding external contoursim, contours, hierarchy = cv2.findContours(fg_mask, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_TC89_L1)# filtering by with, heightfor (i, contour) in enumerate(contours):(x, y, w, h) = cv2.boundingRect(contour)contour_valid = (w >= min_contour_width) and (h >= min_contour_height)if not contour_valid:continue# getting center of the bounding boxcentroid = get_centroid(x, y, w, h)matches.append(((x, y, w, h), centroid))return matches

我們通過高度、寬度和添加質心來添加一些過濾

封裝，算法實現一般流程：

class PipelineRunner(object):'''Very simple pipline.Just run passed processors in order with passing context from one to another.You can also set log level for processors.'''def __init__(self, pipeline=None, log_level=logging.DEBUG):self.pipeline = pipeline or []self.context = {}self.log = logging.getLogger(self.__class__.__name__)self.log.setLevel(log_level)self.log_level = log_levelself.set_log_level()def set_context(self, data):self.context = datadef add(self, processor):if not isinstance(processor, PipelineProcessor):raise Exception('Processor should be an isinstance of PipelineProcessor.')processor.log.setLevel(self.log_level)self.pipeline.append(processor)def remove(self, name):for i, p in enumerate(self.pipeline):if p.__class__.__name__ == name:del self.pipeline[i]return Truereturn Falsedef set_log_level(self):for p in self.pipeline:p.log.setLevel(self.log_level)def run(self):for p in self.pipeline:self.context = p(self.context)self.log.debug("Frame #%d processed.", self.context['frame_number'])return self.contextclass PipelineProcessor(object):'''Base class for processors.'''def __init__(self):self.log = logging.getLogger(self.__class__.__name__)

作為輸入構造函數將采取將按順序運行的處理器列表。每個處理器都是這項工作的一部分。讓我們創建輪廓檢測處理器。

class ContourDetection(PipelineProcessor):'''Detecting moving objects.Purpose of this processor is to subtrac background, get moving objectsand detect them with a cv2.findContours method, and then filter off-bywidth and height. bg_subtractor - background subtractor isinstance.min_contour_width - min bounding rectangle width.min_contour_height - min bounding rectangle height.save_image - if True will save detected objects mask to file.image_dir - where to save images(must exist). '''def __init__(self, bg_subtractor, min_contour_width=35, min_contour_height=35, save_image=False, image_dir='images'):super(ContourDetection, self).__init__()self.bg_subtractor = bg_subtractorself.min_contour_width = min_contour_widthself.min_contour_height = min_contour_heightself.save_image = save_imageself.image_dir = image_dirdef filter_mask(self, img, a=None):'''This filters are hand-picked just based on visual tests'''kernel = cv2.getStructuringElement(cv2.MORPH_ELLIPSE, (2, 2))# Fill any small holesclosing = cv2.morphologyEx(img, cv2.MORPH_CLOSE, kernel)# Remove noiseopening = cv2.morphologyEx(closing, cv2.MORPH_OPEN, kernel)# Dilate to merge adjacent blobsdilation = cv2.dilate(opening, kernel, iterations=2)return dilationdef detect_vehicles(self, fg_mask, context):matches = []# finding external contoursim2, contours, hierarchy = cv2.findContours(fg_mask, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_TC89_L1)for (i, contour) in enumerate(contours):(x, y, w, h) = cv2.boundingRect(contour)contour_valid = (w >= self.min_contour_width) and (h >= self.min_contour_height)if not contour_valid:continuecentroid = utils.get_centroid(x, y, w, h)matches.append(((x, y, w, h), centroid))return matchesdef __call__(self, context):frame = context['frame'].copy()frame_number = context['frame_number']fg_mask = self.bg_subtractor.apply(frame, None, 0.001)# just thresholding valuesfg_mask[fg_mask < 240] = 0fg_mask = self.filter_mask(fg_mask, frame_number)if self.save_image:utils.save_frame(fg_mask, self.image_dir +"/mask_%04d.png" % frame_number, flip=False)context['objects'] = self.detect_vehicles(fg_mask, context)context['fg_mask'] = fg_maskreturn contex

所以只需將BG減法、濾波和檢測部分合并在一起即可。
現在讓我們創建一個處理器，它將鏈接到不同幀上的檢測對象，并創建路徑，也將計算到達出口區域的車輛。

'''Counting vehicles that entered in exit zone.Purpose of this class based on detected object and local cache createobjects pathes and count that entered in exit zone defined by exit masks.exit_masks - list of the exit masks.path_size - max number of points in a path.max_dst - max distance between two points.'''def __init__(self, exit_masks=[], path_size=10, max_dst=30, x_weight=1.0, y_weight=1.0):super(VehicleCounter, self).__init__()self.exit_masks = exit_masksself.vehicle_count = 0self.path_size = path_sizeself.pathes = []self.max_dst = max_dstself.x_weight = x_weightself.y_weight = y_weightdef check_exit(self, point):for exit_mask in self.exit_masks:try:if exit_mask[point[1]][point[0]] == 255:return Trueexcept:return Truereturn Falsedef __call__(self, context):objects = context['objects']context['exit_masks'] = self.exit_maskscontext['pathes'] = self.pathescontext['vehicle_count'] = self.vehicle_countif not objects:return contextpoints = np.array(objects)[:, 0:2]points = points.tolist()# add new points if pathes is emptyif not self.pathes:for match in points:self.pathes.append([match])else:# link new points with old pathes based on minimum distance between# pointsnew_pathes = []for path in self.pathes:_min = 999999_match = Nonefor p in points:if len(path) == 1:# distance from last point to currentd = utils.distance(p[0], path[-1][0])else:# based on 2 prev points predict next point and calculate# distance from predicted next point to currentxn = 2 * path[-1][0][0] - path[-2][0][0]yn = 2 * path[-1][0][1] - path[-2][0][1]d = utils.distance(p[0], (xn, yn),x_weight=self.x_weight,y_weight=self.y_weight)if d < _min:_min = d_match = pif _match and _min <= self.max_dst:points.remove(_match)path.append(_match)new_pathes.append(path)# do not drop path if current frame has no matchesif _match is None:new_pathes.append(path)self.pathes = new_pathes# add new pathesif len(points):for p in points:# do not add points that already should be countedif self.check_exit(p[1]):continueself.pathes.append([p])# save only last N points in pathfor i, _ in enumerate(self.pathes):self.pathes[i] = self.pathes[i][self.path_size * -1:]# count vehicles and drop counted pathes:new_pathes = []for i, path in enumerate(self.pathes):d = path[-2:]if (# need at list two points to countlen(d) >= 2 and# prev point not in exit zonenot self.check_exit(d[0][1]) and# current point in exit zoneself.check_exit(d[1][1]) and# path len is bigger then minself.path_size <= len(path)):self.vehicle_count += 1else:# prevent linking with path that already in exit zoneadd = Truefor p in path:if self.check_exit(p[1]):add = Falsebreakif add:new_pathes.append(path)self.pathes = new_pathescontext['pathes'] = self.pathescontext['objects'] = objectscontext['vehicle_count'] = self.vehicle_countself.log.debug('#VEHICLES FOUND: %s' % self.vehicle_count)return context

?

總結

以上是生活随笔為你收集整理的基于OpenCV的车辆计数（一）的全部內容，希望文章能夠幫你解決所遇到的問題。

如果覺得生活随笔網站內容還不錯，歡迎將生活随笔推薦給好友。