动态物体检测(python)

2022/1/8 12:33:38

本文主要是介绍动态物体检测(python),对大家解决编程问题具有一定的参考价值,需要的程序猿们随着小编来一起学习吧!

简介

本篇博文将实现基于python的运动物体检测。

依赖库

opencv-python

基本动作检测

在计算机视觉中,我们把运动看作是环境的变化。为了计算转换,我们必须有一个背景图像来比较。所以,我们在程序的开头保存第一个图像。

# Converting the image to GrayScale
gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)
gray = cv2.GaussianBlur(gray,(21,21),0)
# Saving the First Frame
if first_frame is None:
    first_frame = gray
    continue

然后,我们将后续帧与保存的第一帧进行比较,以观察差异。计算完差异后,我们可以应用阈值将其转换为黑白图像。

#Calculates difference to detect motion
delta_frame = cv2.absdiff(first_frame, gray)
#Applies Threshold and converts it to black & white image
thresh_delta = cv2.threshold(delta_frame, 30, 255, cv2.THRESH_BINARY)[1]
thresh_delta = cv2.dilate(thresh_delta, None, iterations=0)
#finding contours on the white portion(made by the threshold)
cnts,_ = cv2.findContours(thresh_delta.copy(), cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)

最后一个命令在该黑白图像中查找轮廓,并给出用于创建边界框的坐标,如上面的视频所示。使用运动检测的好处:

  • 它不会保存无用的空闲镜头。因此,减少了其他算法的工作量,因为不会保存空闲帧进行处理。

  • 它需要较少的计算,并且适合实时实施。

加强版

定期更新参照帧

给定的因素导致轮廓检测不理想,运动检测的幼稚方法会在执行开始时为所有比较保存第一帧。不好有几个原因:

  • 白天的照明条件可能会改变。

  • 天气变化。

  • 执行时相机被遮挡。

解决方案:在没有运动的情况下,可以通过定期定期更新保存的帧来轻松解决此问题。

# Number of idle frames to pass before changing the saved frame 
# for further comparisions
FRAMES_TO_PERSIST = 1000

然后将其放在while循环中:

#increment delay counter for every idle frame
delay_counter += 1
#Update the saved first frame
if delay_counter > FRAMES_TO_PERSIST:
    delay_counter = 0
    first_frame = next_frame

过滤微小物体

当检测到运动时,将delay_counter设置为零,微小的物体(例如蜜蜂和昆虫)和通常不必要的轻微不必要的运动被存储起来。解决方案:如片段所示,我们应该在该区域设置一个阈值。

# Minimum boxed area(in pixels) for a detected motion to count as actual motion
# Use to filter out noise or small objects
MIN_SIZE_FOR_MOVEMENT = 2000

然后在while循环中放置一个if语句:

#Checks if the area is big enough to be considered as motion.
if cv2.contourArea(c) > MIN_SIZE_FOR_MOVEMENT:
    #Your code

完整代码

import cv2
import numpy as np
import time

class VideoReader(object):
    def __init__(self, file_name):
        self.file_name = file_name
        try:  # OpenCV needs int to read from webcam
            self.file_name = int(file_name)
        except ValueError:
            pass

    def __iter__(self):
        self.cap = cv2.VideoCapture(self.file_name)
        if not self.cap.isOpened():
            raise IOError('Video {} cannot be opened'.format(self.file_name))
        return self

    def __next__(self):
        was_read, img = self.cap.read()
        if not was_read:
            raise StopIteration
        return img


if __name__ == '__main__':
    frame_provider = VideoReader(0)
    FRAMES_TO_PERSIST = 50
    MIN_SIZE_FOR_MOVEMENT = 200

    delay_counter = 0
    first_frame = None
    delay = 1
    for frame in frame_provider:
        start = time.time()
        # Converting the image to GrayScale
        gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)
        gray = cv2.GaussianBlur(gray, (21, 21), 0)
        # Saving the First Frame
        if first_frame is None:
            first_frame = gray
            continue

        # Calculates difference to detect motion
        delta_frame = cv2.absdiff(first_frame, gray)
        # Applies Threshold and converts it to black & white image
        thresh_delta = cv2.threshold(delta_frame, 30, 255, cv2.THRESH_BINARY)[1]
        thresh_delta = cv2.dilate(thresh_delta, None, iterations=0)
        # finding contours on the white portion(made by the threshold)
        cnts, _ = cv2.findContours(thresh_delta.copy(), cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)

        if np.sum(np.where(thresh_delta > 0, 1, 0)) < MIN_SIZE_FOR_MOVEMENT:
            thresh_delta = np.zeros_like(gray)

        end = time.time()
        seconds = end - start
        # Calculate frames per second
        print(f"seconds: {seconds}")
        fps = 1 / seconds
        imgVis = np.dstack([thresh_delta for _ in range(3)])
        cv2.putText(imgVis, f"FPS: {int(fps)}", (50, 50),
                    cv2.FONT_HERSHEY_COMPLEX, 0.5, (0, 0, 255))
        cv2.imshow('Lightweight Human Pose Estimation Python Demo', imgVis)
        key = cv2.waitKey(delay)
        if key == 27:  # esc
            continue
        elif key == 112:  # 'p'
            if delay == 1:
                delay = 0
            else:
                delay = 1

        # increment delay counter for every idle frame
        delay_counter += 1
        # Update the saved first frame
        if delay_counter > FRAMES_TO_PERSIST:
            delay_counter = 0
            first_frame = gray


这篇关于动态物体检测(python)的文章就介绍到这儿,希望我们推荐的文章对大家有所帮助,也希望大家多多支持为之网!


扫一扫关注最新编程教程