Object Tracking 

-Akshat

Object tracking is the process of locating a moving object (or multiple objects) over time using a camera. It has a variety of uses, some of which are: human-computer interaction, security and surveillance, augmented reality, traffic control, medical imaging and video editing.

Akshat

Some Day2Day Examples

Akshat
It seems interesting
?
Akshat

so for next 80 mins..

a short introduction on video capture

describe the object to be tracked

contour extraction

sketch the frame

Akshat

Some

Results

 

Akshat

before we do anything

import cv2
import numpy as np
Akshat

let' s capture a video

import cv2
import numpy as np

cap = cv2.VideoCapture(0)

while True:
	# Let's capture/read the frames for the video
	_, frame = cap.read() 

	# This frame captured is flipped, Lets's get the mirror effect
	frame = cv2.flip(frame, 1)

	# Any operations on the frame captured will be processed in the operate(frame) method
	final_frame = operate(frame)

	# Showing the captured frame
	cv2.imshow('garrix', frame)
	cv2.imshow('martin', final_frame)

	# Continuous, Large amount of frames produce a video
	# waitkey(value), value is the ammount of millisecs a frame must be displayed
	# 0xff represents the ASCII value for the key, 27 is for ESC
	# waitKey is necessary to show a frame
	if cv2.waitKey(1) & 0xff == 27:
		break


cap.release()
cv2.destroyAllWindows()
Akshat

let' s describe the object

# Info : ZOOM in the frame to get the RGB value of the corresponding pixel

# The range of color for the object to be detected in HSV
hsv_supremum = np.array([172, 221, 255])
hsv_infinum = np.array([150, 40, 130])


def operate(frame):
	frame_copy = frame.copy()
	frame_ret = frame.copy()

	# Converting the frame from BGR format to HSV format
	frame_copy = cv2.cvtColor(frame, cv2.COLOR_BGR2HSV)

	# Thresholding the frame to get our object(color) tracked
	mask = cv2.inRange(frame_copy, hsv_infinum, hsv_supremum)
	mask = cv2.medianBlur(mask, 5)
	mask = cv2.erode(mask, None, iterations=2)
	mask = cv2.dilate(mask, None, iterations=2)

	#cv2.imshow('mask', mask)

	# Obtaining the binary Image for the corresponding mask
	res = cv2.bitwise_and(frame_ret,frame_ret, mask= mask)

	#cv2.imshow('res', res)
	# The result can be improved by smoothening the frame

        # To be contd...
Akshat

my object

Akshat

object configuration

Akshat

object configuration

hsv_supremum = np.array([172, 221, 255])
hsv_infinum = np.array([150, 40, 130])


def operate(frame):
   
    # Some code above
    # Converting the frame from BGR format to HSV format, Hue Saturation Value GOOGle for more
    frame_copy = cv2.cvtColor(frame, cv2.COLOR_BGR2HSV)

    # Thresholding the frame to get our object(color) tracked
    mask = cv2.inRange(frame_copy, hsv_infinum, hsv_supremum)
	
    # Some code below
Akshat

object configuration



# In the operate function

        mask = cv2.medianBlur(mask, 5)
	mask = cv2.erode(mask, None, iterations=2)
	mask = cv2.dilate(mask, None, iterations=2)


	# The result can be further improved by smoothening the frame
        # Issues:
            # Lightning
            # Variable Contour

operations to enhance the object

Akshat

dilation

Akshat

source UDACITY

erosion

Akshat

source UDACITY

Results After Dilation & Erosion in the binary Image

Akshat

contour extraction

# In the same operate function

        mask_copy = mask.copy()
	cnts = cv2.findContours(mask.copy(), cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)[-2]
	center = None

	if len(cnts) > 0:
		c = max(cnts, key = cv2.contourArea)
		((x, y), radius) = cv2.minEnclosingCircle(c)

		M = cv2.moments(c)
		center = (int(M["m10"] / M["m00"]), int(M["m01"] / M["m00"]))

		if radius > 10:
			cv2.circle(frame_ret, (int(x), int(y)), int(radius), (0, 150, 255), 2)
			cv2.circle(frame_ret, center, 5, (0, 0, 255), -1)

operations to extract the contour

Akshat

Object Contour Extracted

Object Mask Extracted

Video Capture

Akshat

sketch the frame

import cv2
import numpy as np 

from collections import deque           # <--

pts = deque()                           # <--
cap = cv2.VideoCapture(0)

# In the operate function

        pts.appendleft(center)
	#print(pts[0])

	for i in range(1 , len(pts)):
		if pts[i-1] is None or pts[i] is None:
			continue
		thickness = 6
		cv2.line(frame_ret, pts[i-1], pts[i], (0, 0, 255), thickness)


	return frame_ret

operations to sketch on the frame

Akshat

Sample Sketch

Akshat

Sample Sketch

Akshat

Code

Implementation

Lets Begin

[Quest]ions

?

Thank you

 

Akshat Sharma

Sophomore

 

Phoenix | CV Last Lecture

By Akshat Sharma

Made with Slides.com

Phoenix | CV Last Lecture

Project Based session on Computer Vision, covering Object Tracking & Webcam Drawing.

  • 388

More from Akshat Sharma