Drawing grid navigation by opencv map reading

Realization effect

Get a map .jpg As shown in the figure , The final effect is MATLAB Draw a grid map in

utilize PC Image preprocessing with the drawing tool

for example ： Acquired pictures test.jpg The pixel size displayed in the drawing software is 412162, Import to MATLAB The number of grids is also 412162 It's not reasonable , So resize the picture in the drawing software , Make a pixel correspond to a grid point , Here it is scaled to a size of 40*15 A grid matrix .

utilize OpenCV Recognize the image and get the obstacle information

OpenCV Procedure steps for identifying obstacle information ：

1、 Image Reading
2、 Replace the picture with a binary diagram
3、 Get pixel information （ What we get here is the index value ）
4、 Store data .txt File is used to MATLAB Read

OpenCV Specific procedures for reading drawings , It is worth noting that , The test image itself has two colors , Black and white , If additional information needs to be identified , utilize OpenCV Threshold processing of , To generate the binary graph you need .

import cv2    

#  Read the picture 
img = cv2.imread('test.jpg') 

#  Binary image processing 
img_gray = cv2.cvtColor(img,cv2.COLOR_BGR2GRAY)
ret,img_bin = cv2.threshold(img_gray, 127, 255, cv2.THRESH_BINARY)

#  Check whether the processed image meets the requirements 
cv2.imshow("img_bin", img_bin) 
cv2.waitKey(0)
cv2.destroyAllWindows()

'''
 Capture pixels of the image , among 0 It means black ,255 It means white 
row and col Represents the rows and columns of the image , Pixels [row,col] The value of the pixel corresponding to its position 
image_line  Store linear value pictures ,1 It means an obstacle ,0 Indicates a free zone 
image_index Store the index value of picture pixels , Reference resources MATLAB Of ind2sub Function interpretation 
'''

#  Get the pixel row and column height of the picture 
row,col = img_bin.shape

#  Define data types 
image_line = []
index = 0
image_index = []


#  Circular judgment of pixel information 
for c in range(col):
    for r in range(row):
        index = index + 1
        if img_bin[r,c] == 0:
            image_line.append(1)
            image_index.append(index)
        else:
            image_line.append(0)
    image_line.append('^p')

#  Write map data to txt file 
file = open('data_map.txt','w') 
file.write(str(image_line)); 
file.close()

#  Write index data to txt file 
file = open('data_index.txt','w') 
file.write(str(image_index)); 
file.close()

Image pixels row and col The explanation of

Picture in OpenCV The point relation corresponding to each pixel in the is ：

View image data

open data_map.txt Copied to the Word in , In the replacement function , take ’^p’ Replace with ^p Implement line break , Copy data to txt Better viewing effect ;
take data_index.txt The beginning of the data [ And the end ] Symbol deletion
data_map The effect of data ：

take index Data import MATLAB Draw grid

Will go to the beginning [ And the end ] Of data_index Data import MATLAB

Follow the procedure below to read index Data path planning , Path planning reference link : 5 MATLAB And grid path planning algorithm （1 Rasterize images and dijkstra Algorithm ）.

clc;
clear;
data = load('data.txt');

XRows = 40; YCols = 15;
StartCoord = [3,6];
GoalCoord = [35,2];

[obsx,obsy] = ind2sub([YCols,XRows],data);
ObsCoord = [obsy' obsx'];

[Opt,Field] = Dijkstra_main(XRows,YCols,StartCoord,GoalCoord,ObsCoord);

Achieve the effect of path planning ：

Summary of this chapter

1、 Grid map in MATLAB The degree of reproduction depends on the pixels , So it is not necessary to modify the picture size in the drawing software ,OpenCV Get row and col The map can also be reproduced according to the bisection relationship
2、 Picture maps and MATLAB The map of has an inverse relationship , You can consider using the drawing software to reverse the picture vertically before reading the picture