Setting a path in a web program that uses Python and OpenCV

[InJae1213]

Select Topic Area

question

Body

I'm going to use YOLOv4 to create a program that takes pictures of the desired object when it comes into the camera angle.
Below is the content of the code app.py for that.

from flask import Flask, request, jsonify
from flask_cors import CORS
import cv2
import numpy as np
import base64

app = Flask(__name__)
CORS(app)  # Enable CORS for all routes

YOLO_CFG =  "C:/Users/shini/webcam-app/webcam_server/yolov4.cfg"
YOLO_WEIGHTS = "C:/Users/shini/webcam-app/webcam_server/yolov4.weights"
YOLO_CLASSES = "C:/Users/shini/webcam-app/webcam_server/coco.names"
CONFIDENCE_THRESHOLD = 0.5  # adjust to your needs
NMS_THRESHOLD = 0.4  # adjust to your needs

net = cv2.dnn.readNet(YOLO_WEIGHTS, YOLO_CFG)
layer_names = net.getLayerNames()
output_layers = [layer_names[i - 1] for i in net.getUnconnectedOutLayers().flatten()]
classes = open(YOLO_CLASSES).read().strip().split("\n")

def detect_objects(image):
 height, width, channels = image.shape
 blob = cv2.dnn.blobFromImage(image, 0.00392, (416, 416), (0, 0, 0), True, crop=False)
 net.setInput(blob)
 outs = net.forward(output_layers)
 class_ids = []
 confidences = []
 boxes = []
 for out in outs:
     for detection in out:
         scores = detection[5:]
         class_id = np.argmax(scores)
         confidence = scores[class_id]
         if confidence > CONFIDENCE_THRESHOLD:
             center_x = int(detection[0] * width)
             center_y = int(detection[1] * height)
             w = int(detection[2] * width)
             h = int(detection[3] * height)

             x = int(center_x - w / 2)
             y = int(center_y - h / 2)

             boxes.append([x, y, w, h])
             confidences.append(float(confidence))
             class_ids.append(class_id)

 detected_objects = []
 indexes = cv2.dnn.NMSBoxes(boxes, confidences, CONFIDENCE_THRESHOLD, NMS_THRESHOLD)
 for i in range(len(boxes)):
     if i in indexes:
         detected_objects.append(classes[class_ids[i]])

 return detected_objects

@app.route('/detect', methods=['POST'])
def detect():
 image_data = request.form.get("image")
 if not image_data:
     return jsonify(error="no image provided"), 400

 image_data = base64.b64decode(image_data.split(",")[1])
 image = cv2.imdecode(np.frombuffer(image_data, dtype=np.uint8), cv2.IMREAD_COLOR)

 detected_objects = detect_objects(image)
 return jsonify(detected_objects=detected_objects)

if __name__ == '__main__':
 app.run(host='0.0.0.0', port=5000)

I modified it to a relative route to distribute this program.

YOLO_CFG = "yolov4.cfg"
YOLO_WEIGHTS = "yolov4.weights"
YOLO_CLASSES = "coco.names"
CONFIDENCE_THRESHOLD = 0.5  # adjust to your needs
NMS_THRESHOLD = 0.4  # adjust to your needs

But there was an error

OpenCV(4.8.1) D:\a\opencv-python\opencv-python\opencv\modules\dnn\src\darknet\darknet_importer.cpp:210: error: (-212:Parsing error) Failed to open NetParameter file: yolov4.cfg in function 'cv::dnn::dnn4_v20230620::readNetFromDarknet'
File "C:\Users\shini\new-webcam-app\webcam_server\app.py", line 17, in
net = cv2.dnn.readNet(YOLO_WEIGHTS, YOLO_CFG)

However, if v3 is used instead of YOLOv4, no error has occurred even when using the relative path.
I wonder how YOLOv4 is used and set as a relative path does not cause an error. I ask for your help me.

Guidelines

• I have read and understood this category's guidelines before making this post.

[rantirules]

Make sure that your yolov4.cfg, yolov4.weights, and coco.names files all in the same directory as your app.py file.
But in addition to that specify the full file paths like you did initially instead of using a relative path.
You might also need to change the working directory in your script using:

import os

# Get the current working directory
cwd = os.getcwd()

# Change the working directory to where the files are located
os.chdir(os.path.join(cwd, 'yolov4_files'))

# And then now you should be able to use the relative paths for your YOLO file paths
YOLO_CFG = "yolov4.cfg"
YOLO_WEIGHTS = "yolov4.weights"
YOLO_CLASSES = "coco.names"

[ChilledFerrum]

In reply to what rantirules mentioned, make sure you are using the right path for your data files.

It appears that your code produces an error at the function readNetFromDarknet.
A fix for this could be to parse your config file first and then your weights file, as seen in OpenCV's readNetFromDarknet function.

Here's a fix that could work for you better

cv2::dnn::readNetFromDarknet(YOLO_CFG, YOLO_WEIGHTS)

Hope this helps! Have a great day