0:00 opencv is an open source Library
0:02 consisting of hundreds of computer
0:03 vision algorithms while the library was
0:05 mainly developed in C and C plus it can
0:08 be used in other languages such as Java
0:10 JavaScript R and of course python opencv
0:13 python is a collection of python
0:14 bindings which allow data scientists and
0:16 Engineers to use opencv using a python
0:18 interface the core operations module
0:20 contains the basic building blocks of
0:22 opencv such as data structures and
0:24 procedures specific to computer vision
0:26 applications such as pixel editing
0:28 geomatic Transformations and code
0:30 optimization the core functionality
0:32 includes reading and editing individual
0:34 pixels or regions of interest within
0:36 images splitting and merging channels
0:38 padding images matte operations and
0:41 images such as addition blending
0:43 operations or bitwise operations and
0:46 last but not least functionality to
0:47 measure performance and optimize opencv
0:50 code image processing with opencv can be
0:52 used to change the color space of an
0:54 image for example from blue green red to
0:56 grayscale or HSV Hue saturation value
0:59 color model and to track a colored
1:02 object in an image or in a video to
1:04 perform geometric Transformations on
1:06 images such as scaling translation
1:08 rotation a fine transformation and
1:11 perspective transformation convert
1:13 images to Binary images using image
1:15 thresholding to smoothen blur and filter
1:17 images to apply morphological
1:19 Transformations on the shapes of images
1:21 like erosion dilatation opening closing
1:24 morphological gradient top hat or black
1:27 hat to find image gradients using Sobel
1:30 and laplacian derivatives to perform
1:31 Edge detection using the Kani Edge
1:34 detection algorithm to create image
1:35 pyramids on multiple levels as well as
1:38 using pyramids for image blending to be
1:40 able to find and draw Contours to be
1:43 able to find image histograms to plot
1:45 image histograms as well as to analyze
1:47 image histograms to use the Fourier
1:49 transform of an image in order to
1:51 perform Edge detection to be able to use
1:53 the template matching method to detect
1:55 either one or multiple matching objects
1:58 in an image to use the whole line
2:00 transform in order to detect lines in an
2:03 image or the whole circle transform to
2:05 detect circles in an image using the
2:07 Watershed algorithm in order to segment
2:10 images using the grab cut algorithm to
2:13 extract the foreground of an image
2:14 feature detection functionalities in
2:16 opencv help us to better understand the
2:19 features of a given image to use the
2:21 Harris Corner detection algorithm to
2:23 detect Corners in images or the shy
2:25 tomasi Corner detector for tracking
2:27 Corners as well as low scale invariant
2:30 feature transform detector for Corners
2:33 in images with changing scales as well
2:35 as its faster Alternatives the speeded
2:37 up robust features detector the fast
2:39 detector for real-time applications the
2:42 brief detector outputting binary
2:44 descriptors the free and unpatented orb
2:47 detector the feature detection module in
2:49 opencv also has functionality to match
2:51 features across different images video
2:54 analysis and opencv includes techniques
2:56 such as main shift cam shift and Optical
2:59 flow minship finds the area of Maximum
3:01 pixel density in an image by placing a
3:04 window at a random position over an
3:06 image and then Shifting the window
3:07 during repeated iterations until it
3:09 converges to the area of Maximum density
3:11 mean shift uses a fixed size window
3:13 regardless of image which can be an
3:15 issue for some applications camshaft
3:17 relies on applying mean shift first then
3:19 it fits a scaled rotated ellipse
3:21 detected area after which minshift is
3:24 run again using the new ellipse window
3:26 Optical flow is the pattern of apparent
3:28 motion of image objects between the
3:30 consecutive frames of a video and can be
3:33 detected using the Lucas can 8 method
3:35 implemented in opencv camera calibration
3:37 and 3D reconstruction can be performed
3:40 as well in opencv opencv helps you
3:42 identify distortions caused by cameras
3:44 in images such as radial distortions
3:47 when straight lines appear curved or
3:49 tangential Distortion when some areas of
3:51 the image look nearer than expected
3:53 opencv can help to detect the object or
3:55 image Corner points and use them as
3:57 input for the camera calibration
3:58 procedure to underscore or the original
4:00 image finally opencv can also help to
4:03 estimate the reprojection error after
4:05 calibration opencv can be used for pose
4:07 estimation to understand how an object
4:09 is situated in space and once that is
4:11 achieved we can render 3D objects within
4:13 the image itself opencv can also
4:15 estimate depth between images using
4:17 epipolar geometry as well as stereo
4:20 images Machine learning in opencv is
4:22 implemented using several models such as
4:24 K nearest neighbors and support Vector
4:27 machines for classification and
4:28 handwritten data extraction as well as
4:30 k-means clustering for applications such
4:33 as color quantization that is reducing
4:35 the number of colors in an image in
4:37 computational photography techniques in
4:39 opencv include image denoising to remove
4:41 noise using non-local means denoising
4:44 image in painting to restore old
4:46 degraded images affected by black spots
4:49 and strokes and lastly to generate and
4:51 display High dynamic range images or HDR
4:54 images from multiple exposed images and
4:56 use the exposure Fusion technique to
4:58 merge our exposure sequence object
5:00 detection is one of the main
5:01 applications of computer vision and it's
5:03 implemented in opencv using the Cascade
5:05 classifier opencv provides the Cascade
5:08 classifiers using the har feature-based
5:10 object detection which can detect eyes
5:12 faces or other objects in images or in
5:15 video streams opencv bindings allow data
5:18 scientists to leverage and extend the
5:20 opencv C plus modules with python if you
5:23 want me to make more advanced videos on
5:25 opencv please leave a comment down below
5:27 subscribe to the channel and hit the
5:29 notification Bell to stay up to date
5:31 with my future videos
Chrome Extension