OpenCV getting error "Cant parse contours. Input argument doesnt provide sequence protocol" in Python ROS

Question

I’m working on a line detection function for my robot. I tried to combine 2 different pieces of code, because the first program didn’t work with ros.

I’m getting the error in the cv2.drawContours() function and I don’t know how to solve this problem. I have tried Pascal’s solution to switch c with contours, but it didn’t help.

How can I make c work in the drawcontours function?

Here is my code:

#! /usr/bin/env python
import rospy
import cv2
import numpy as np
from sensor_msgs.msg import Image
from sensor_msgs.msg import CompressedImage
from cv_bridge import CvBridge, CvBridgeError
import traceback
import sys

class LineDetector:
    def __init__(self):
        self.rate = rospy.Rate(rospy.get_param("/rate/lineDetector"))
        self.bridge = CvBridge()
        self.image_sub = rospy.Subscriber("/camera/image_raw", Image, self.image_callback_raw)
        self.image_sub_rpi =rospy.Subscriber("/raspicam_node/image/compressed", CompressedImage, self.image_callback_compressed)
        self.image_pub = rospy.Publisher("processed_image", Image, queue_size=1)
        self.low_b = np.uint8([5, 5, 5])
        self.high_b = np.uint8([0, 0, 0])
        self.c = 0
    
    def image_callback_raw(self, msg):
        try:
            cv_image = self.bridge.imgmsg_to_cv2(msg, "passthrough")
            self.process_image(cv_image)
        except Exception as e:
            traceback.print_exc()
            rospy.logerr(e)
            rospy.logerr("CvBridge Error, skipped image frame!")

    def image_callback_compressed(self, msg):
        try:
            np_arr = np.frombuffer(msg.data, np.uint8)
            image_np = cv2.imdecode(np_arr, cv2.IMREAD_COLOR)
            self.process_image(image_np)
        except Exception as e:
            rospy.logerr(e)
            rospy.logerr("skipped processed image frame!")
        
    def process_image(self, cv_image):
        mask = cv2.inRange(cv_image, self.low_b, self.high_b)
        contours, hierarchy = cv2.findContours(mask, 1, cv2.CHAIN_APPROX_NONE)
        if len(contours) > 0:
            self.c = max(contours, key=cv2.contourArea)
            M = cv2.moments(self.c)
            if M["m00"] != 0:
                cx = int(M["m10"] / M['m00'])
                cy = int(M["m01"] / M["m00"])
                print("CX: " + str(cx) + "  CY: " + str(cy))
                if cx >= 120:
                    print("Turn Left")
                if cx < 120 and cx > 40:
                    print ("On Track")
                if cx <= 40:
                    print("Turn Right")
                cv2.circle(cv_image, (cx,cy), 5, (255,255,255), -1)
        else:
            print("I dont see the line")
        cv2.drawContours(cv_image, self.c, -1, (0, 255, 0), -1)
        image_message = self.bridge.cv2_to_imgmsg(cv_image, "passthrough")
        self.image_pub.publish(image_message)
    
    def run(self):
        while not rospy.is_shutdown():
            self.rate.sleep()    

if __name__ == '__main__':
    rospy.init_node('line_detector')
    line_detector = LineDetector()
    rospy.spin()

Asked By: Cooler Mann

||

Source

Answer 1

The drawContours() function should only be called when you actually have detected the contours (when len(contours)>0):

    def process_image(self, cv_image):
        print("processing")
        mask=cv2.inRange(cv_image,self.low_b,self.high_b)
        # I suggest you to visualize your mask
        #cv2.imshow("mask",mask)
        #cv2.waitKey(1)
        contours, _ =cv2.findContours(mask, 1, cv2.CHAIN_APPROX_NONE) # hierarchy is redundant
        if len(contours)>0:
            self.c = max(contours, key=cv2.contourArea)
            M = cv2.moments(self.c)
            if M["m00"] != 0:
                cx = int(M["m10"]/M['m00'])
                cy = int(M["m01"]/M["m00"])
                print("CX: "+str(cx)+"  CY: "+str(cy))
                if cx >= 120:
                    print("Turn Left")
                if cx < 120 and cx > 40:
                    print ("On Track")
                if cx <=40:
                    print("Turn Right")
                cv2.circle(cv_image, (cx,cy), 5, (255,255,255), -1)
            
            # we can draw contours only in case that they were detected
            cv2.drawContours(cv_image, self.c, -1,(0,255,0),-1) 
        else:
            print("I dont see the line")
        image_message = self.bridge.cv2_to_imgmsg(cv_image, "rgb8") # fixed encoding
        self.image_pub.publish(image_message)

Also, please note that you probably also need to change the passthrough encoding to rbg8 to enable the conversion of the image to ROS Image msg.

Nevertheless, I suppose that another problem lies in the RGB thresholds (low_b and high_b) that you have set. These values obviously depend on your particular case, but I recommend you set them somewhat higher:

        self.low_b =np.uint8([0,0,0]) 
        self.high_b=np.uint8([100,100,100])

(It is also better to visualize your mask for debug proposes, as I have mentioned in the code snippet above).
This is the visualization of the resulting image_message:

Answered By: Stefan

OpenCV getting error "Cant parse contours. Input argument doesnt provide sequence protocol" in Python ROS

Question:

Answers: