Line Follower with OpenCV first lesson

Follower_CV_1.py

from picamera.array import PiRGBArray
from picamera import PiCamera
import time
import cv2
import numpy as np
import RPi.GPIO as GPIO

GPIO.setmode(GPIO.BOARD)
GPIO.setup(40, GPIO.OUT)
GPIO.output(40, GPIO.HIGH)

camera = PiCamera()
camera.resolution = (640, 360)
camera.rotation = 180
rawCapture = PiRGBArray(camera, size=(640, 360))
time.sleep(0.1)

for frame in camera.capture_continuous(rawCapture, format="bgr", use_video_port=True):	
	image = frame.array
	roi = image[200:250, 0:639]
	Blackline= cv2.inRange(roi, (0,0,0), (50,50,50))
	kernel = np.ones((3,3), np.uint8)
	Blackline = cv2.erode(Blackline, kernel, iterations=5)
	Blackline = cv2.dilate(Blackline, kernel, iterations=9)	
	img,contours, hierarchy = cv2.findContours(Blackline.copy(),cv2.RETR_TREE,cv2.CHAIN_APPROX_SIMPLE)	
	if len(contours) > 0 :
	   x,y,w,h = cv2.boundingRect(contours[0])	   
	   cv2.line(image, (x+(w/2), 200), (x+(w/2), 250),(255,0,0),3)
	cv2.imshow("orginal with line", image)	
	rawCapture.truncate(0)	
	key = cv2.waitKey(1) & 0xFF	
	if key == ord("q"):
		break

GPIO.output(40, GPIO.LOW)

The angle works something like the D in a PID controller. The angle tells you the trend.