Can I use the visualization code for result

[moonseokha]

Thank you so much for sharing the code.
I would like to use the result visualization in HiVT code, but has the visualization code shown on the github page not been shared? Can you share it ?
Thank U!

[ZikangZhou]

Hi @moonseokha,

Sorry for the late reply.

def plot_single_vehicle(
    avm: ArgoverseMap, 
    sample_past_trajectory: np.ndarray,  # (20, 2)
    sample_groundtruth: np.ndarray,  # (20, 2) 
    sample_forecasted_trajectories: List[np.ndarray],  # List[(30, 2)]
    sample_city_name: str, 
    ls: str):

    sample_groundtruth = np.concatenate((np.expand_dims(sample_past_trajectory[-1], axis=0), sample_groundtruth), axis=0)
    for sample_forecasted_trajectory in sample_forecasted_trajectories:
        sample_forecasted_trajectory = np.concatenate((np.expand_dims(sample_past_trajectory[-1], axis=0), sample_forecasted_trajectory), axis=0)

    ## Plot history
    obs_len = sample_past_trajectory.shape[0]
    pred_len = sample_groundtruth.shape[0]
    plt.plot(
        sample_past_trajectory[:, 0],
        sample_past_trajectory[:, 1],
        color="#ECA154",
        label="Past Trajectory",
        alpha=1,
        linewidth=3,
        zorder=15,
        ls = ls
    )

    ## Plot future
    plt.plot(
        sample_groundtruth[:, 0],
        sample_groundtruth[:, 1],
        color="#d33e4c",
        label="Ground Truth",
        alpha=1,
        linewidth=3.0,
        zorder=20,
        ls = "--"
    )

    ## Plot prediction
    for j in range(len(sample_forecasted_trajectories)):
        plt.plot(
            sample_forecasted_trajectories[j][:, 0],
            sample_forecasted_trajectories[j][:, 1],
            color="#007672",
            label="Forecasted Trajectory",
            alpha=1,
            linewidth=3.2,
            zorder=15,
            ls = "--"
        )
        
        # Plot the end marker for forcasted trajectories
        plt.arrow(
            sample_forecasted_trajectories[j][-2, 0], 
            sample_forecasted_trajectories[j][-2, 1],
            sample_forecasted_trajectories[j][-1, 0] - sample_forecasted_trajectories[j][-2, 0],
            sample_forecasted_trajectories[j][-1, 1] - sample_forecasted_trajectories[j][-2, 1],
            color="#007672",
            label="Forecasted Trajectory",
            alpha=1,
            linewidth=3.2,
            zorder=15,
            head_width=1.1,
        )
    
    ## Plot the end marker for history
    plt.arrow(
            sample_past_trajectory[-2, 0], 
            sample_past_trajectory[-2, 1],
            sample_past_trajectory[-1, 0] - sample_past_trajectory[-2, 0],
            sample_past_trajectory[-1, 1] - sample_past_trajectory[-2, 1],
            color="#ECA154",
            label="Past Trajectory",
            alpha=1,
            linewidth=3,
            zorder=25,
            head_width=1.0,
        )

    ## Plot the end marker for future
    plt.arrow(
            sample_groundtruth[-2, 0], 
            sample_groundtruth[-2, 1],
            sample_groundtruth[-1, 0] - sample_groundtruth[-2, 0],
            sample_groundtruth[-1, 1] - sample_groundtruth[-2, 1],
            color="#d33e4c",
            label="Ground Truth",
            alpha=1,
            linewidth=3.0,
            zorder=25,
            head_width=1.0,
        )

    ## Plot history context
    for j in range(obs_len):
        lane_ids = avm.get_lane_ids_in_xy_bbox(
            sample_past_trajectory[j, 0],
            sample_past_trajectory[j, 1],
            sample_city_name,
            query_search_range_manhattan=65,
        )
        [avm.draw_lane(lane_id, sample_city_name) for lane_id in lane_ids]

    ## Plot future context
    for j in range(pred_len):
        lane_ids = avm.get_lane_ids_in_xy_bbox(
            sample_groundtruth[j, 0],
            sample_groundtruth[j, 1],
            sample_city_name,
            query_search_range_manhattan=65,
        )
        [avm.draw_lane(lane_id, sample_city_name) for lane_id in lane_ids]

[moonseokha]

Thank you for reply and good code. :)

[xinfu607]

Thank you for reply and good code. :)
Hi,
Could you please release the complete visualization code? Thanks a lot.

[Joonki0111]

Hi, i've seen the code and i want to know how to get those samples in the code below

def plot_single_vehicle(
avm: ArgoverseMap,
sample_past_trajectory: np.ndarray, # (20, 2)
sample_groundtruth: np.ndarray, # (20, 2)
sample_forecasted_trajectories: List[np.ndarray], # List[(30, 2)]
sample_city_name: str,
ls: str):

there is no code generating sample_past_trajectory, sample_groundtruth, sample_forecasted_trajectories
in HiVT repository so i cant run that code.

any help and suggestions will be helping my project alot.
Thanks :)

[MPCheng-ZW]

Hi, have you solved this problem? Thank you!
@Joonki0111 @xinfu607 @moonseokha @ZikangZhou

Hi, i've seen the code and i want to know how to get those samples in the code below

def plot_single_vehicle( avm: ArgoverseMap, sample_past_trajectory: np.ndarray, # (20, 2) sample_groundtruth: np.ndarray, # (20, 2) sample_forecasted_trajectories: List[np.ndarray], # List[(30, 2)] sample_city_name: str, ls: str):

there is no code generating sample_past_trajectory, sample_groundtruth, sample_forecasted_trajectories in HiVT repository so i cant run that code.

any help and suggestions will be helping my project alot. Thanks :)

[yaobaishen]

Hi, have you solved this problem? Thank you! @Joonki0111 @xinfu607 @moonseokha @ZikangZhou

Hi, i've seen the code and i want to know how to get those samples in the code below
def plot_single_vehicle( avm: ArgoverseMap, sample_past_trajectory: np.ndarray, # (20, 2) sample_groundtruth: np.ndarray, # (20, 2) sample_forecasted_trajectories: List[np.ndarray], # List[(30, 2)] sample_city_name: str, ls: str):
there is no code generating sample_past_trajectory, sample_groundtruth, sample_forecasted_trajectories in HiVT repository so i cant run that code.
any help and suggestions will be helping my project alot. Thanks :)

Has anyone solved the problem? It would be much better if full visualization code is available.

@ZikangZhou sorry for interruption, but could you look at this issue again? thanks.

[alfredgu001324]

The visualization instructions here also let you save the trajectory as pkl files for later visualization. Some modifications are needed (such as get rid of the uncertainty component) but the transformation is the same.

[yaobaishen]

The visualization instructions here also let you save the trajectory as pkl files for later visualization. Some modifications are needed (such as get rid of the uncertainty component) but the transformation is the same.

@alfredgu001324 thanks for your work! it really saves my day!