Kanade Lucas Tomasi (KLT) optical flow algorithm in Python

Question

Hello. I searched all over the Internet and couldn’t find.

Someone has an alternative to perform feature extraction on images with the python Kanade Lucas Tomasi (KLT) algorithm?

Answer 1

Opencv works with this (or variants of it) algorithm.

Sparse Optical flow: These Algorithms, like the Kanade-Lucas-Tomashi (KLT) Feature tracker, track the Location of a few Feature points in an image.

The example creates a simple app that tracks some points in a video. To decide the points the function is used cv.goodFeaturesToTrack(), Take the first frame, some points (in the corners) are detected with the Shi-Tomasi Corner Detector function, then these points are tracked iteratively through the Lucas-Kanade optical flow. The previous frame and stitches, and the next frame are passed to the function cv.calcOpticalFlowPyrLK(). Return the next points and some status numbers that can be 1 if the next point is found, if not zero.

See the code:

import numpy as np
import cv2 as cv
cap = cv.VideoCapture('slow.flv')
# params for ShiTomasi corner detection
feature_params = dict( maxCorners = 100,
                       qualityLevel = 0.3,
                       minDistance = 7,
                       blockSize = 7 )
# Parameters for lucas kanade optical flow
lk_params = dict( winSize  = (15,15),
                  maxLevel = 2,
                  criteria = (cv.TERM_CRITERIA_EPS | cv.TERM_CRITERIA_COUNT, 10, 0.03))
# Create some random colors
color = np.random.randint(0,255,(100,3))
# Take first frame and find corners in it
ret, old_frame = cap.read()
old_gray = cv.cvtColor(old_frame, cv.COLOR_BGR2GRAY)
p0 = cv.goodFeaturesToTrack(old_gray, mask = None, **feature_params)
# Create a mask image for drawing purposes
mask = np.zeros_like(old_frame)
while(1):
    ret,frame = cap.read()
    frame_gray = cv.cvtColor(frame, cv.COLOR_BGR2GRAY)
    # calculate optical flow
    p1, st, err = cv.calcOpticalFlowPyrLK(old_gray, frame_gray, p0, None, **lk_params)
    # Select good points
    good_new = p1[st==1]
    good_old = p0[st==1]
    # draw the tracks
    for i,(new,old) in enumerate(zip(good_new,good_old)):
        a,b = new.ravel()
        c,d = old.ravel()
        mask = cv.line(mask, (a,b),(c,d), color[i].tolist(), 2)
        frame = cv.circle(frame,(a,b),5,color[i].tolist(),-1)
    img = cv.add(frame,mask)
    cv.imshow('frame',img)
    k = cv.waitKey(30) & 0xff
    if k == 27:
        break
    # Now update the previous frame and previous points
    old_gray = frame_gray.copy()
    p0 = good_new.reshape(-1,1,2)
cv.destroyAllWindows()
cap.release()

Upshot:

Source: https://docs.opencv.org/3.4/d7/d8b/tutorial_py_lucas_kanade.html