image_graph_line_to_digital_convert

results

note

the digital form of the points and the preview line graph (like above) will be stored automatically.

usage

set a few parameters args in run_allimgs.m, also some de_noise related in the head of imgPlot2digital.m.

linemover mode

There're two modes of removing the grid lines. By setting the formal parameter linemover of the function imgPlot2digital(), results are shown in above.

• The scan mode is done by setting the step_x and step_y vars, so when the grid line is covered by the plot line, it'll remove the plot line while removing the grid line.
• The imclose mode is a digital image processing skill, when the plot line is bolder than the grid line, after binarizing the image, the black pixel size in the gird line is smaller than its in the plot line, so by setting the structuring element size we can remove the small black pixels, aka the grid line.
• Empty char [] stands DO NOT apply grid lines removing algorithm.

what the program does

1. convert the image to gray, then binary.
2. read the "black" pixels and convert them from the image coordinate system to the "coordinate"(the axis in the image) coordinate system.
3. flip the coordinate system, bc matlab use -y in image coordinate system.
4. drop some points (like remove the grid lines in axis, see in function reduceLines, reduce_xy and de_noiser).
5. make x-y pairs unique. now we have a perfect line plot~
6. do an interpret from the points we have above to obtain the points at x-axis u set.
7. find the first derivative of the points above and get rid of the values bigger than a threshold (set the formal parameter of function final_plot_filter())
8. smooth filter the points above using mean filter.
9. DONE!

run

simply run run_allimgs.m and wait for the magic happens.

If u find this repo helps, a star will be the best support.