问题
I am working with a sensor and I have collected data which represented in two lists:
filtered_x = [64, 90, 83, 78, 57, 58....]
filtered_y = [26, 17, 63, 21, 62, 86....]
Which gives the following scatter plot:
This is part of a calibration process. In order to complete this, I have to find the circles in the plot and come up with coordinates of the centers of the circles so the sensor can be calibrated. Which libraries should I use and how do I go about doing this? I have come across nearest K neighbor but I cannot find any practical approaches to applying this to both lists to identify the centers of the 25 clusters. The readings as a result of a flashing LED board.
回答1:
Use kernel density estimation (KDE), and find maxima.
Because that is what you want to find, the most dense locations.
来源：https://stackoverflow.com/questions/44658086/finding-circles-in-scatter