MORP is a novel algorithm for small robots, designed to drive efficiently in an indoor environment with little information to refer to. It renders a virtual lanes on the ground and detect objects upon the lanes, and further avoid it. In this project, we only use depth camera 'Intel D435i' for object handling, path planning and driving.
import pyrealsense2 as rs
import numpy as np
import cv2
import matplotlib.pyplot as plt
import math
import time #if you want to see FPS
import easyGo #if you want to move the robot
go to source ./ground_seg.py
#if ground_seg.py is in the same folder
import ground_segdef verticalGround(depth_image2, images, numCol, plot):
return images, dead_endIt vertically scans depth image to segment ground. Return painted color image and the last pixel's y-value. If argument 'plot' is True, you can see the matplot graph of current column.
def GroundSeg(depth_image, color_image, stride=160):
return temp_image, virtual_lane_availableIt uses 'VerticalGround()' to widely segment ground. Return widely painted color image, and a list of 'dead_end' with 160 pixel stride - 'virtual_lane_available'. It will be used for path planning.
def LaneHandling(virtual_lane_available, unavailable_thres, n):
return direcIt uses dead_ends of each lanes for path planning. Return 'direc' means which direction to go. Finally, you can drive the robot with GoEasy(direc).
