[ACCEPTED]-One dimensional edge detection-edge-detection

Accepted answer

Score: 10

One way to get to your desired result is 17 to adapt the 2D Canny edge detector as follows 16 (code in Mathematica):

First, compute the 15 spatial derivative using a Gaussian derivative 14 filter, setting the sigma value relative 13 to the scale of the edges you want to detect. Take 12 the absolute value of the result.

d = Abs@GaussianFilter[data, {{10, 5}}, 1];

Then, determine 11 a threshold automatically to cluster the 10 previous derivative values in two groups 9 (here using Otsu's method).

thrd = FindThreshold[d];

Then, detect 8 the steps of the derivative values (transitions 7 into/from the "dead band").

steps = Flatten@Image`StepDetect[d, thrd]["NonzeroPositions"];

At this point 6 you have the ends of the edges:

ListLinePlot[data, Epilog -> {Red, PointSize[Large], Map[Point[{#, data[[#]]}] &, steps]}]

enter image description here

Optionally--it 5 seems that's what you'd like--keep only 4 the lowest ends of the edges. Clustering 3 the data points at the ends of the edges 2 works in this case, but I'm not sure how 1 robust it is.

t = FindThreshold@data[[steps]];
steps2 = Select[steps, data[[#]] <= t &];

ListLinePlot[data, Epilog -> {Red, PointSize[Large], Map[Point[{#, data[[#]]}] &, steps2]}]

enter image description here

Score: 2

Given the nice contrast of these edges, there 9 is an easy solution that will work robustly: detect 8 all monotonous sequences of pixel values 7 (strictly increasing or decreasing). You 6 will keep the sequences having a total height 5 above a threshold (50 in your case) to reject 4 the noisy peaks.

As a byproduct, you'll get 3 the starting and ending points (not exactly 2 where you expect them though, but this can 1 be improved on if needed).

Barcodes ?

Score: 1

So you are looking for a particular change 5 in slope - ie a certain change in Y per 4 sample?

Isn't it simply look at the difference 3 in Y between two samples and if it's absolute 2 value changed by more than some limit mark 1 that as an edge?

Source: stackoverflow.com

[ACCEPTED]-One dimensional edge detection-edge-detection

Edge Detection method better than Canny Edge detection

Detect an object in a camera image in C#

How detect long edges of wall to prepare mask and recolor

Efficient way to combine intersecting bounding rectangles

Find checkbox contours using OpenCV

How to warp the later frame to previous frame using optical flow image

How to remove small object in image with Python

Remove background text and noise from an image using image processing with OpenCV

Adjusting and Improving X-ray image's contrast and its quality

How to detect the Sun from the space sky in OpenCv?