Algorithm 1. Valid data segment fusion and partitioning algorithm

Input: valid_data_segments, start_indexs, end_indexs Output: cars_list, front_edges, back_edges, occ_type_list 1. for each data segment i: 2.  if cars_list is empty: 3.    create a new car segment 4.  else: 5.    let j be the last car segment, and l be the second-to-last segment (if it exists) 6.    if the minimum difference between segment i and segment l < $T_e$ and time difference < $P_f$: 7.      fuse with segment l 8.    else if the minimum difference between segment i and segment j < $T_e$ and time difference < $P_f$: 9.      fuse with segment j 10.  else if the time difference < $P_o$: 11.    mark as occlusion 12.  else: 13.    create a new car segment End