Panorama(=Surround) view

1. Creating 360 degree Panoramic Views

• Project the points in 3D space into cylindrical surface

• LiDAR센서의 특징에 따라 설정값이 달라 진다.

• h_res: Horizontal resolution
• v_res: vertical resolution

# KTTI dataset = Velodyne HDL 64E
## A vertical field of view of 26.9 degrees, at a resolution of 0.4 degree intervals. The vertical field of view is broken up into +2 degrees above the sensor, and -24.9 degrees below the sensor.
## A horizontal field of view of 360 degrees, at a resolution of 0.08 - 0.35 (depending on the rotation rate)
## Rotation rate can be selected to be betwen 5-20Hz.
## http://velodynelidar.com/docs/datasheet/63-9194%20Rev-E_HDL-64E_S3_Spec%20Sheet_Web.pdf

# Resolution and Field of View of LIDAR sensor
h_res = 0.35 # horizontal resolution, assuming rate of 20Hz is used
v_res = 0.4 # vertical res
v_fov = (-24.9, 2.0) # Field of view (-ve, +ve) along vertical axis
v_fov_total = -v_fov[0] + v_fov[1]

2. Surround View

SEED = 202

import math
import random
import numpy as np
random.seed(SEED)
np.random.seed(SEED)

import cv2
from lidar import *


##   cylindrial projection
SURROUND_U_STEP = 1.    #resolution
SURROUND_V_STEP = 1.33
SURROUND_U_MIN, SURROUND_U_MAX = np.array([0,    360])/SURROUND_U_STEP  # horizontal of cylindrial projection
SURROUND_V_MIN, SURROUND_V_MAX = np.array([-90,   90])/SURROUND_V_STEP  # vertical   of cylindrial projection


def lidar_to_surround(lidar):
    def normalise_to_255(a):
        return (((a - min(a)) / float(max(a) - min(a))) * 255).astype(np.uint8)

    x = lidar['x']
    y = lidar['y']
    z = lidar['z']
    r = lidar['intensity']
    d = np.sqrt(x ** 2 + y ** 2)  # map distance relative to origin
    u,v = lidar_to_surround_coords(x,y,z,d)

    width  = int(SURROUND_U_MAX - SURROUND_U_MIN + 1)
    height = int(SURROUND_V_MAX - SURROUND_V_MIN + 1)
    surround     = np.zeros((height, width, 3), dtype=np.float32)
    surround_img = np.zeros((height, width, 3), dtype=np.uint8)

    surround[v, u, 0] = d
    surround[v, u, 1] = z
    surround[v, u, 2] = r
    surround_img[v, u, 0] = normalise_to_255(np.clip(d,     0, 30))
    surround_img[v, u, 1] = normalise_to_255(np.clip(z+1.8, 0, 100))
    surround_img[v, u, 2] = normalise_to_255(np.clip(r,     0, 30))

    return surround, surround_img

def lidar_to_surround_coords(x, y, z, d ):
    u =   np.arctan2(x, y)/np.pi*180 /SURROUND_U_STEP
    v = - np.arctan2(z, d)/np.pi*180 /SURROUND_V_STEP
    u = (u +90)%360  ##<todo> car will be spit into 2 at boundary  ...

    u = np.rint(u)
    v = np.rint(v)
    u = (u - SURROUND_U_MIN).astype(np.uint8)
    v = (v - SURROUND_V_MIN).astype(np.uint8)

    return u,v

lidar = np.load("/root/share/project/didi/data/didi/didi-2/Data/1/15/lidar/1530509304325762000.npy")
surround, surround_img = lidar_to_surround(lidar)
cv2.imwrite("./output/surround.png",surround_img)
from IPython.display import Image
Image(filename="./output/surround.png")

• Ronny Restrepo

windowsub0406/KITTI_Tutorial

Velodyne -> Panoramic Image: Convert Velodyne data(model : HDL-64E) to panoramic image.