added notebooks

nan removal

best mean=1.7695378948997886

point_radius=45,

split some code into functions point_radius=40,

a lot of changes, I want to see how it performs (rewrote evaluation to load data in batches)

pointcloud_depth_coeff=1.05

dist_coeff=0.1

clustering_eps=80

clustering_eps=150

added dbscan

added clustering and merge_th=100

some depth maps are nov retrieved from the pointcloud

added missed edge detection, but it is not much better

scale_estimation_coefficient=4.5, added hip, reordered operations to remove loops

slightly improved performance

half implemented missing point detection; scale_estimation_coefficient =4

gridsearch parameters

entries become suddenly empty

algorithm update: in order to be fit, the detected line is required to be in the same direction as proposed one

most strict for the night

stricter edge selsection, but there must must be a rule that the line must not deviate a couple degrees from detected when fitting to edges

line stuff, now full hough, liberal edge selection

remove self loops and bi-directional non unique edges. edge_th=60

reverted color range change

fixed bad logic in line selection, vectorized the calculations, added hough lines, changed distance to sqeuclidean, rewrote convert_entry_to_human_readable to use match case

added cleaner vertex mask

Update handcrafted_solution.py

Update handcrafted_solution.py

added line-component separation using vertex masks

included flashing

improved apex detection algorithm

Upload 2 files