Finding a fixed number of interpolation points between acquired points

I have some points acquired on google earth to be able to trace a circuit in matlab, in particular the goal is to identify the centerline starting from the side carriageways. Being a hand selection of points, the two carriageways acquired have a number of different points and therefore it is impossible to find a centerline by taking the average values of the coordinates. So I tried to use the 'interparc' function that allows you to create a number of points chosen from those initially fixed, the problem is that the points thus created are not equidistant, since the points fixed at the start were not.

So I was wondering if there is a method to create a fixed number of interpolating points, all equally spaced, among those initially available.

[Matt J]

Examples?

[RAFFAELE MAGLIONE]

I drew a path from Google Earth pro, exported the kml file and converted it to excel. Then imported into matlab, so I have two coordinate vectors (Latitude and Longitude) of for example 500 points and 600 points (the internal carriageway has fewer points since it encloses a smaller area). The problem is that I have data acquired from a vehicle's GPS that identify its lap on the track with about 7000 points.

The 'interparc' function allows you to transform the 500 and 600 points of the first two vectors into 7000 points by interpolation, but these 7000 points are not equidistant since on the straight the points acquired by Google Earth are further away than those on the curve.

This means that when I go to find the centerline by averaging the coordinates of the two interpolated vectors, this centerline does not actually come out in the center.

The only solution I found was to manually make the number of points of the two initial vectors the same (e.g. 600) on Google Earth, so I could calculate the centerline directly by averaging these two vectors and subsequently interpolating them, but I it seems a 'non-automatic' method.

[Jan]

The currently provided information do not help to solve the problem. For Matlab it does not matter, if the data are coordinates from Google Earth or the price of bananas. Instead of explaining the physical meaning, post some example data in the way they are represented in your case. Maybe a simple rand() command is sufficient already, maybe in the combination with sort().

What does this mean mathematically: "trace a circuit in matlab, in particular the goal is to identify the centerline starting from the side carriageways."

[RAFFAELE MAGLIONE]

I try to explain better:

I start with Coord_int and Coor_est of different dimensions and after using 'interparc' I obtain Coord_int_interp and Coord_est_interp of same dimension:

Then I calculate Coord_cen through the mean value of Latitude and Longitude of Coord_int_interp and Coord_est_interp.

Graphically this are the results:

In the curve the red line (that is the centerline) isn't at the center of the track.

[Bruno Luong]

Can you provide Coord_est and Coord_int ?

[Mathieu NOE]

hello @RAFFAELE MAGLIONE

no big surprise here, your method is not a true centerline definition

a better approach involves to compute the normal vector at each point of one curve and find the corresponding point of the other curve according to this (normal) vector. The mid point of this segment is your centerline

see for example :

How to find the centreline of two curves - (mathworks.com)

all the best

[RAFFAELE MAGLIONE]

Thank you @Mathieu NOE, it worked well

[Mathieu NOE]

My pleasure !