Opencv error 215 assertion failed

preface

The author started to learn OpenCV these days. Looking at the Demo made by the official document, he found that he encountered an error. The prompt is as follows:

error: (-215:Assertion failed) !empty() in function ‘cv::CascadeClassifier::detectMultiScale’

Screenshot in PyCharm

Note: the following will begin to write the author’s method to solve the error step by step (fei hua). If the reader wants to solve it immediately, he can directly jump to the fourth part and change the reference to the original path of the reader

1. Cause of error

The original code that generated the error

# Import camera package
import cv2

import numpy as np

# Load face detector, eye detector, smile detector
# The source of the error code!!!!!!!!!!!!!!!!!
face_cascade = cv2.CascadeClassifier('haarcascade_files/haarcascade_frontalface_default.xml')
# face_cascade = cv2.CascadeClassifier(pathf)

eye_cascade = cv2.CascadeClassifier(pathe)

smile_cascade = cv2.CascadeClassifier(paths)

# Get camera
# If there are multiple cameras, you can change 0 to 1, 2 and 3
cap = cv2.VideoCapture(0)

# loop
while (True):
    # Get the picture taken by the camera
    ret, frame = cap.read()
    
    # !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
    faces = face_cascade.detectMultiScale(frame, 1.3, 2)	#An error occurred in this code
    # !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
    
    img = frame

Error: (- 215: assertion failed)! The main reason for empty() in function ‘cv::CascadeClassifier::detectMultiScale’ is that our code can not find the correct calling face classifier correctly, in other words, it does not find the correct file path

This error will also occur if the author randomly enters a nonexistent file path, so it further verifies my idea:)

In other words, as long as our call path is correct, we can run normally, so our main consideration is to change the code segment of the import detection package, that is

# Load face detector, eye detector, smile detector
face_cascade = cv2.CascadeClassifier('haarcascade_files/haarcascade_frontalface_default.xml')

2. Tried method

Several wrong writing methods are introduced (the author’s configuration is win10, python + Anaconda + Python 3.6 environment. If the author’s configuration is unsuccessful, it does not mean that your xdm configuration is unsuccessful. It is for reference only

1. Native writing of official documents

face_cascade = cv2.CascadeClassifier('haarcascade_files/haarcascade_frontalface_default.xml')

1. Writing of single slash

face_cascade=cv2.CascadeClassifier('D:/opencv/opencv/data/haarcascades/haarcascade_frontalface_default.xml')

1. Plus. load

face_cascade.load('D:/opencv/opencv/sources/data/haarcascades/haarcascade_frontalface_default.xml')

After trying these methods, they all report the same error, indicating that the path reference is still wrong

3. Find path

First, we need to find the path where the classifier is located. We can enter it directly in Cmd

pip show opencv-python

To find and display the installation path of OpenCV

It should be noted that the path of OpenCV installed by Anaconda here is not placed with the original installation package. You need to use another way to view it on anaconda. You need to find the directory package with cv as the file name. The following is the location directory of that package

This is the file directory, which corresponds to the package we want to reference

4. Final solution

Change the index to: computer absolute path + double slash

The corresponding code is:

pathf = 'E:\EnvironmentsA\cv\opencv_python-3.4.5.20-cp36-cp36m-win_amd64\cv2\data\haarcascade_frontalface_default.xml'

# Load face detector, eye detector, smile detector
face_cascade = cv2.CascadeClassifier(pathf)

Then there will be no error when calling below

 # Get the picture taken by the camera
    ret, frame = cap.read()
    faces = face_cascade.detectMultiScale(frame, 1.3, 2)
    img = frame

5. Test source code

Finally, the source code for testing is attached

# -*- coding:utf-8 -8-
"""
Author: FeverTwice
Date: 2021--08--28
"""

# Import camera package
import cv2

import numpy as np

pathf = 'E:\EnvironmentsA\cv\opencv_python-3.4.5.20-cp36-cp36m-win_amd64\cv2\data\haarcascade_frontalface_default.xml'
pathe = 'E:\EnvironmentsA\cv\opencv_python-3.4.5.20-cp36-cp36m-win_amd64\cv2\data\haarcascade_eye.xml'
paths = 'E:\EnvironmentsA\cv\opencv_python-3.4.5.20-cp36-cp36m-win_amd64\cv2\data\haarcascade_smile.xml'

# Load face detector, eye detector, smile detector
# face_cascade = cv2.CascadeClassifier('haarcascade_files/haarcascade_frontalface_default.xml')
face_cascade = cv2.CascadeClassifier(pathf)

eye_cascade = cv2.CascadeClassifier(pathe)

smile_cascade = cv2.CascadeClassifier(paths)

# Get camera
# If there are multiple cameras, you can change 0 to 1, 2 and 3
cap = cv2.VideoCapture(0)

# loop
while (True):
    # Get the picture taken by the camera
    ret, frame = cap.read()
    faces = face_cascade.detectMultiScale(frame, 1.3, 2)
    img = frame

    for (x, y, w, h) in faces:
        # Draw the face, blue, brush width
        img = cv2.rectangle(img, (x, y), (x + w, y + h), (255, 0, 0), 2)

        # Draw the face area and detect the human eye in the face area to save computing resources
        face_area = img[y:y + h, x:x + w]

        ## Human eye detection
        # The human eye cascade classifier engine is used to recognize the face region, and the returned eyes are the eye coordinates
        eyes = eye_cascade.detectMultiScale(face_area, 1.3, 10)

        for (ex, ey, ew, eh) in eyes:
            # Draw the human eye socket, green, with a width of 1
            cv2.rectangle(face_area, (ex, ey), (ex + ew, ey + eh), (0, 255, 0), 1)

        ## Smiley face detection
        #  The smile cascade classifier engine is used for human eye recognition in the region, and the returned is the eye coordinates
        smiles = smile_cascade.detectMultiScale(face_area, scaleFactor=1.16, minNeighbors=65, minSize=(25, 25),
                                                flags=cv2.CASCADE_SCALE_IMAGE)

        for (ex, ey, ew, eh) in smiles:
            # Draw the smile box, the color is red, and the brush width is 1
            cv2.rectangle(face_area, (ex, ey), (ex + ew, ey + eh), (0, 0, 255), 1)
            cv2.putText(img, 'Smile', (x, y - 7), 3, 1.2, (0, 0, 255), 2, cv2.LINE_AA)

    # Display the effect screen
    cv2.imshow('frame2', img)

    # Didn't 5ms monitor a keyboard action
    if cv2.waitKey(5) & 0xFF == ord('q'):
        break

# Turn off the camera
cap.release()
# Close the image window
cv2.destroyAllWindows()

Posted by Abner
at Aug 31, 2021 — 2:47 AM
Tag:
OpenCV
Pycharm
Python
Programming
Computer Vision