Revision history [back]

I was reading through the OpenCV tutorial for image scanning when I ran into a peculiar line of code (link):

// accept only char type matrices
CV_Assert(I.depth() != sizeof(uchar));

This code partially works because I.depth() returns CV_8U (which is 0), while sizeof(uchar) returns 1. If an array of another type is passed in though, the code will fail unless (coincidentally) I.depth() returns 1.

Here is what I recommend replacing these lines with:

// accept only char type matrices
CV_Assert(I.depth() == CV_8U);

I was reading through the OpenCV tutorial for image scanning when I ran into a peculiar line of code (link):

// accept only char type matrices
CV_Assert(I.depth() != sizeof(uchar));

This code partially works because I.depth() returns CV_8U (which is 0), while sizeof(uchar) returns 1. If an array of another type is passed in though, the code will fail unless (coincidentally) I.depth() returns 1.

Here is what I recommend replacing these lines with:

// accept only char type matrices
CV_Assert(I.depth() == CV_8U);

I was reading through the OpenCV tutorial for image scanning when I ran into a peculiar line of code (link):

// accept only char type matrices
CV_Assert(I.depth() != sizeof(uchar));

This code partially works because I.depth() returns CV_8U (which is 0), while sizeof(uchar) returns 1. If an array of another type is passed in though, the code will fail unless (coincidentally) I.depth() returns 1.

Here is what I recommend replacing these lines with:

// accept only char type matrices
CV_Assert(I.depth() == CV_8U);