Individual Report
Due date: 23 Oct 2024
Include Code, figures
Submit: PDF file only to Blackboard
Q&A: Contact at the Bottom of this web page
Questions Image Resources: Link
Give the codes and answers to the following questions about Numpy
, use only one line code
How to find the maximum value of each row in a 2-dimensional NumPy array? (Find the maximum value of each row in a given array)
np.random.seed(100)
a = np.random.randint(1,10, [5,3])
#> array([[9, 9, 4],
#> [8, 8, 1],
#> [5, 3, 6],
#> [3, 3, 3],
#> [2, 1, 9]])
How to compute the result of dividing the min-by-max of each row of a 2-dimensional NumPy array? (Given a 2-dimensional NumPy array, compute min-by-max for each row)
np.random.seed(100)
a = np.random.randint(1,10, [5,3])
a
#> array([[9, 9, 4],
#> [8, 8, 1],
#> [5, 3, 6],
#> [3, 3, 3],
#> [2, 1, 9]])
How to remove all missing values from a NumPy array? (Remove all nan values from a 1-dimensional NumPy array)
# input
np.array([1,2,3,np.nan,5,6,7,np.nan])
# expect output
array([ 1., 2., 3., 5., 6., 7.])
How to find all local maxima (peaks) in a 1-dimensional array? (Find all the peaks in a 1-dimensional array a. A peak is a number that is larger than the numbers on either side of it.)
# input
a = np.array([1, 3, 7, 1, 2, 6, 0, 1])
# expect output
array([2, 5])
How to invert all columns of a 2D array? (Reverse all columns in 2D array arr)
# input
arr = np.arange(9).reshape(3,3)
Calculate and report the camera's intrinsics and distortion coefficient of the provided images on a checkerboard (The checkerboards square actual size is 20x20 mm. The images of checkerboards are in the folder Q2_calibration
).
Using the provided set of 9 images (The images are in the folder Q3_stitching
):
Implement a stereo matching algorithm to generate a disparity map from the provided pair of stereo images (The images are in the folder Q4_disparity
).
Leaf Contour Extraction: Extract the contour of the green leaf in the image (The image in the folder Q5_contour
), and discuss potential challenges and solutions for extracting contours from more complex backgrounds than the simple single-colored desktop.