Visualization

[derkaczda]

Hey, thank you for the great paper and code! Does there exist code to visualize the predictions? I couldn't find it.

Thanks in advance!

[xinge008]

We use Open3D to visulize the results and the color map is shown in the yaml;

Another alternative is to utilize the visualization method shown in SECOND repo.

[derkaczda]

Thanks for the quick response! For the Open3D approach do you have code published for this project?

[OuyangJunyuan]

Thanks for the quick response! For the Open3D approach do you have code published for this project? I just have a simple one for viz. you need it?

[derkaczda]

Thanks @OuyangJunyuan, I would like to have the code for your simple viz. You can send it via mail [email protected]. Thank you very much :)

[Carl12138aka]

Thanks @OuyangJunyuan, I would like to have the code for your simple viz. You can send it via mail [email protected]. Thank you very much :)

hi, could you please share it with me? thanks a lot! my email is : [email protected]

[OuyangJunyuan]

Thanks @OuyangJunyuan, I would like to have the code for your simple viz. You can send it via mail [email protected]. Thank you very much :)

hi, could you please share it with me? thanks a lot! my email is : [email protected]

of course~

[gwag00]

Thanks for the quick response! For the Open3D approach do you have code published for this project? I just have a simple one for viz. you need it?

hi, Sorry for bother you.. but could you send it to me? my email is : [email protected]

[artofstate]

@OuyangJunyuan hi,could you please share it with me ? My email is [email protected] Thank you very much!!!

[TimGor1997]

Thanks for the quick response! For the Open3D approach do you have code published for this project? I just have a simple one for viz. you need it?

hi, could you please share the vis code with me? My email is [email protected] Thank you very much!!!

[TimGor1997]

Thanks @OuyangJunyuan, I would like to have the code for your simple viz. You can send it via mail [email protected]. Thank you very much :)

hi, could you please share the vis code with me? My email is [email protected] Thank you very much!!!

[Kin-Zhang]

If possible, could you please post the visualization code on the issue for us to do quick visual with Open3D

[derkaczda]

I already wrote @OuyangJunyuan about that couple of weeks ago but did not get any reply, I don't want to post his code without his consent.

[OuyangJunyuan]

I already wrote @OuyangJunyuan about that couple of weeks ago but did not get any reply, I don't want to post his code without his consent.

oh！it's rarely to check my email. I will show it flowing ,just very simple and crude one：

import numpy as np
import yaml
import open3d
from pathlib import Path

 
points_dir = Path('/home/ou/Documents/dataset/my_pvrcnn_dataset(gazebo)/kitti/object/training/velodyne') # path to .bin data
label_dir = Path('/home/ou/workspace/code/Cylinder3D-master/my_result') # path to .label data

label_filter = [40, 48, 70, 72]    # object's label which you wan't to show
# label_filter = []
with open('/home/ou/workspace/code/Cylinder3D-master/config/label_mapping/semantic-kitti.yaml', 'r') as stream: # label_mapping configuration file
    label_mapping = yaml.safe_load(stream)
    color_dict = label_mapping['color_map']


def get_rgb_list(_label):
    c = color_dict[_label]
    return np.array((c[2], c[1], c[0]))


def draw_pc(pc_xyzrgb):
    pc = open3d.geometry.PointCloud()
    pc.points = open3d.utility.Vector3dVector(pc_xyzrgb[:, 0:3])
    pc.colors = open3d.utility.Vector3dVector(pc_xyzrgb[:, 3:6] / 255.)

    def custom_draw_geometry_with_key_callback(pcd):
        def change_background_to_black(vis):
            opt = vis.get_render_option()
            opt.background_color = np.asarray([0, 0, 0])
            opt.point_size = 1
            return False

        key_to_callback = {}
        key_to_callback[ord("K")] = change_background_to_black
        open3d.visualization.draw_geometries_with_key_callbacks([pcd], key_to_callback)

    custom_draw_geometry_with_key_callback(pc)


def concate_color(_points, _label):
    color = np.zeros((_points.shape[0], 3))
    label_id = np.unique(_label)
    for cls in label_id:
        if label_filter.__len__() == 0:
            color[_label == cls] = get_rgb_list(cls)
        elif label_filter.count(cls) == 0:
            color[_label == cls] = get_rgb_list(cls)
    _points = np.concatenate([_points, color], axis=1)
    return _points


for it in label_dir.iterdir():
    label_file = it
    points_file = points_dir / (str(it.stem) + '.bin')
    label = np.fromfile(label_file, dtype=np.uint32)
    points = np.fromfile(points_file, dtype=np.float32).reshape((-1, 4))[:, 0:3]
    print(label.shape, points.shape)
    colorful_points = concate_color(points, label)
    draw_pc(colorful_points)

[TimGor1997]

Dear Junyuan,

Glad to hear from you! Thanks for your kindly reply!

Your help is greatly appreciated!

Best regards,

Tim

------------------ 原始邮件 ------------------ 发件人: "xinge008/Cylinder3D" @.>; 发送时间: 2021年11月6日(星期六) 中午12:54 @.>; @.@.>; 主题: Re: [xinge008/Cylinder3D] Visualization (#73)

I already wrote @OuyangJunyuan about that couple of weeks ago but did not get any reply, I don't want to post his code without his consent.

oh！it's rarely to check my email. I will show it flowing ,just very simple and crude one： import numpy as np import yaml import open3d from pathlib import Path points_dir = Path('/home/ou/Documents/dataset/my_pvrcnn_dataset(gazebo)/kitti/object/training/velodyne') # path to .bin data label_dir = Path('/home/ou/workspace/code/Cylinder3D-master/my_result') # path to .label data label_filter = [40, 48, 70, 72] # object's label which you wan't to show # label_filter = [] with open('/home/ou/workspace/code/Cylinder3D-master/config/label_mapping/semantic-kitti.yaml', 'r') as stream: # label_mapping configuration file label_mapping = yaml.safe_load(stream) color_dict = label_mapping['color_map'] def get_rgb_list(_label): c = color_dict[_label] return np.array((c[2], c[1], c[0])) def draw_pc(pc_xyzrgb): pc = open3d.geometry.PointCloud() pc.points = open3d.utility.Vector3dVector(pc_xyzrgb[:, 0:3]) pc.colors = open3d.utility.Vector3dVector(pc_xyzrgb[:, 3:6] / 255.) def custom_draw_geometry_with_key_callback(pcd): def change_background_to_black(vis): opt = vis.get_render_option() opt.background_color = np.asarray([0, 0, 0]) opt.point_size = 1 return False key_to_callback = {} key_to_callback[ord("K")] = change_background_to_black open3d.visualization.draw_geometries_with_key_callbacks([pcd], key_to_callback) custom_draw_geometry_with_key_callback(pc) def concate_color(_points, _label): color = np.zeros((_points.shape[0], 3)) label_id = np.unique(_label) for cls in label_id: if label_filter.len() == 0: color[_label == cls] = get_rgb_list(cls) elif label_filter.count(cls) == 0: color[_label == cls] = get_rgb_list(cls) _points = np.concatenate([_points, color], axis=1) return _points for it in label_dir.iterdir(): label_file = it points_file = points_dir / (str(it.stem) + '.bin') label = np.fromfile(label_file, dtype=np.uint32) points = np.fromfile(points_file, dtype=np.float32).reshape((-1, 4))[:, 0:3] print(label.shape, points.shape) colorful_points = concate_color(points, label) draw_pc(colorful_points)

— You are receiving this because you commented. Reply to this email directly, view it on GitHub, or unsubscribe. Triage notifications on the go with GitHub Mobile for iOS or Android.

[zhaodejian]

几周前我已经写过@OuyangJunyuan，但没有得到任何回复，我不想在未经他同意的情况下发布他的代码。

哦！很少查看我的电子邮件。我会展示它的流动性，只是非常简单粗暴的一个：

import  numpy  as  np 
import  yaml 
import  open3d 
from  pathlib  import  Path

 
points_dir  =  Path ( '/home/ou/Documents/dataset/my_pvrcnn_dataset(gazebo)/kitti/object/training/velodyne' ) # .bin 数据的路径
label_dir  =  Path ( '/home/ou/workspace/code/Cylinder3D- master/my_result' ) # .label 数据的路径

label_filter  = [ 40 , 48 , 70 , 72 ]     # 你不想显示的对象标签
# label_filter = [] 
with  open ( '/home/ou/workspace/code/Cylinder3D-master/config/label_mapping/semantic-kitti .yaml' , 'r' ) as  stream : # label_mapping 配置文件
    label_mapping  =  yaml . 安全加载（流）
     color_dict  =  label_mapping [ 'color_map' ]


def  get_rgb_list ( _label ):
     c  =  color_dict [ _label ]
    返回 np。数组(( c [ 2 ], c [ 1 ], c [ 0 ]))


def  draw_pc ( pc_xyzrgb ):
     pc  =  open3d。几何。点云（）
    个人电脑。点 =  open3d。实用程序。Vector3dVector ( pc_xyzrgb [:, 0 : 3 ])
    个人计算机。颜色 =  open3d。实用程序。Vector3dVector ( pc_xyzrgb [:, 3 : 6 ] /  255. )

    def  custom_draw_geometry_with_key_callback ( pcd ):
         def  change_background_to_black ( vis ):
             opt  =  vis。get_render_option ()
            选择。背景颜色 =  np。asarray ([ 0 , 0 , 0 ])
            选择。point_size  =  1
            返回 False

        key_to_callback  = {}
         key_to_callback [ ord ( "K" )] =  change_background_to_black 
        open3d。可视化。draw_geometries_with_key_callbacks ([ pcd ], key_to_callback )

    custom_draw_geometry_with_key_callback ( pc )


def  concate_color ( _points , _label ):
     color  =  np。零(( _points . shape [ 0 ], 3 ))
     label_id  =  np。label_id中cls的唯一（_label）
     ：
        如果label_filter。__len__ () == 0 :
            颜色[ _label == cls ] =        get_rgb_list ( cls )
         elif  label_filter。计数( cls ) ==  0 :
            颜色[ _label  ==  cls ] =  get_rgb_list ( cls )
     _points  =  np。连接([ _points , color ], axis = 1 )
    返回 _points


 在label_dir中为它 。iterdir ():
     label_file = it points_file = points_dir / ( str ( it.stem ) + ' .bin ' ) label =
     np . fromfile ( label_file , dtype = np . uint32 )
    点= np . fromfile ( points_file , dtype = np . float32   
            ）。reshape (( - 1 , 4 ))[:, 0 : 3 ]
     print ( label . shape , points . shape )
     colourful_points  =  concate_color ( points , label )
     draw_pc ( colourful_points )

Is it a point cloud with only one color?

[zhaodejian]

@OuyangJunyuan