CODEWITHSUNDEEP
pythonnumpyopencv
Max
Max

Reputation: 87

OpenCv raising error when numpy array fed into it

I'm making a mandelbrot set zoom video maker and openCV is not allowing me to feed in my numpy array.

My end goal is to make a video that zooms in on "interesting" points in the mandelbrot set automaticly, but for now all I'm doing is zooming in on the center.

This is my code for making the video.

from mandelbrot import getMandelbrot
import numpy as np
from cv2 import VideoWriter, VideoWriter_fourcc
import numpy
width = 500
height = 500
fps = 45
seconds = 1

fourcc = VideoWriter_fourcc(*'MP42')
video = VideoWriter('./mandelzoom.avi', fourcc, float(fps), (width, height))
for i in range(fps*seconds):
    frame = getMandelbrot((-1/(i/(fps*seconds)+1), -1/(i/(fps*seconds)+1), 1/(i/(fps*seconds)+1), 1/(i/(fps*seconds)+1)), (width, height), 75)
    video.write(np.array(frame, dtype=np.int8))
    print("frame "+str(i)+" out of "+str(fps*seconds)+".")
video.release()

This is the mandelbrot module I have made.

import colorsys
def hsv2rgb(h,s,v):
    return [round(i * 255) for i in colorsys.hsv_to_rgb(h,s,v)]
def doesConverge(c, iterations, radius=2):
    x=0
    xpre=0
    for i in range(iterations):
        xpre=x
        x=x*x+c
        if abs(x)>radius:
            return hsv2rgb(i/iterations, 1, 1)
        if xpre==x:
            return [0, 0, 0]
    return [0, 0, 0]
def getMandelbrot(locationRect, size, iterations):
    scaleX = size[0]/abs(locationRect[0]-locationRect[2])
    scaleY = size[1]/abs(locationRect[1]-locationRect[3])
    mandelbrot=[]
    for i in range(size[1]):
        mandelbrot.append([])
        for j in range(size[0]):
            mandelbrot[i].append(doesConverge(complex(j/scaleX+locationRect[0], i/scaleY+locationRect[1]), iterations))
    return mandelbrot

When I run my video maker, it throws this error:

  File "C:/Users/Max/AppData/Local/Programs/Python/Python37/mandelzoom.py", line 15, in <module>
    video.write(np.array(frame, dtype=np.int8))
cv2.error: OpenCV(4.1.0) C:\projects\opencv-python\opencv\modules\videoio\src\cap_ffmpeg.cpp:298: error: (-215:Assertion failed) image.depth() == CV_8U in function 'cv::`anonymous-namespace'::CvVideoWriter_FFMPEG_proxy::write'

Upvotes: 2

Views: 211

Answers (1)

Michael
Michael

Reputation: 2414

opencv is expecting an 8-bit unsigned type (CV_8U, which can contain values 0 to 255) but you're passing an 8-bit signed type (np.int8, which can contain values -128 to 127). Try using np.uint8 instead of np.int8 for the array type

Upvotes: 3

Related Questions