EgalYue
commented
6 years ago Following is a simple model of detected region of marker. I cut this detected region of marker into 3 parts(Area 1, Area 2, Area 3). And for different height(X_w) I test the path based on the following 6 cases(6 Arrows).
+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Firstly, I set the number of robot step as same! In the following figure, each point means each step, different paths are presented with different colors.
Fixed path: The way-points the robot should go
Real path: The way-points the robot actually walks along
Measured path: The way-points which are measured according to camera pose
------------------RESULTS for the same number of steps from A to B !!!-------------------------
We can find that, if the path closer to the marker, the smaller the error is.
Next, I will do more simulations for different number of steps, compare more paths in different cases.....still working on
raultron
I assume that there is a 3m x 6m area and the marker is located at the position as the following figure shown. The detected area of marker is like a semicircle(radius = 3m, blue line) . The robot(red points) move in the detection area of marker. This rectangle area can be converted into a 30 × 60 grid, the length of each cell of grid is 0.1m. This can be stored in a 30 x 60 matrix.
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ Every time i set the _homographyiters = 1000, which means i compute the mean value of each point (camera position in the world coordinate) for each simulation: Just like the figure shows, I computed 5 points(cam positions) for testing: _p1 = p1_sum / 1000 p2 = p2_sum / 1000 p3 = p3_sum / 1000 p4 = p4_sum / 1000 p5 = p5sum / 1000
The following show the results of 2 simulations, for every simulation we got a really different results!!!
1. simulation:
-- Fix path--: [[ 0.45 0.45 0.45 0.45 0.45] [ 1.65 1.75 1.85 1.95 2.05]] -- Measured path --: [[ 0.45 0.44961373 0.45022644 0.44889577 0.45027423] [ 1.65 1.75180091 1.85093572 1.95183743 2.05137137]] -- Real path--: [[ 0.45 0.45 0.45038627 0.45015982 0.45126406] [ 1.65 1.75 1.84819909 1.94726337 2.04542594]]
-- Fix_path --: [[ 0.45 0.45 0.45 0.45 0.45] [ 1.65 1.75 1.85 1.95 2.05]] -- Measured path --: [[ 0.45 0.44943768 0.44863969 0.44859178 0.44952944] [ 1.65 1.75104356 1.85008045 1.94938169 2.05039602]] -- Real path--: [[ 0.45 0.45 0.45056232 0.45192263 0.45333086] [ 1.65 1.75 1.84895644 1.94887599 2.04949431]]