[Python] How can i detect the objects by using a custom-trained RoboFlow data?

I have been working on the detection of the objects based on the my custom-trained RoboFlow data (not yolov8n.pt). My code can open the webcam and take a photo, then it can classify the object by using the created image.

It works; however, my code overwrites the new image and processes it again. I tried to use real-time webcam data to perform the same task, unfortunately i do not know how to achieve this.

May you please help me about this issue?

Thank you,

import roboflow
import cv2

rf = roboflow.Roboflow(api_key="XYZ")
project = rf.workspace().project("ABC")
model = project.version(1).model

cap = cv2.VideoCapture(0)
cam_control = cap.isOpened()

prev_frame_time = 0

while cam_control == True:
ret,frame = cap.read()
if ret == False:
print("Image not found!")
break

cv2.imwrite("captured_image.png", frame)

prediction = model.predict("captured_image.png").json()


I have tried cv2.imshow, cv2.imread etc., but i could not solve the problem.

Continue reading...