Some Bonsai.Vision operators that use OpenCV functions have constraints on the input IplImage that are not documented in the operator description or XML doc comments.
For example, the GoodFeaturesToTrack operator only accepts an 8-bit or 32 bit, single channel image.
The resulting error is hard to parse. For instance, using this operator on a normal RGB camera image results in:
Assertion failed: src.type() == CV_8UC1 || src.type() == CV_32FC1 in function cv::cornerEigenValsVecs, corner.cpp(263)
It would be good to document these somewhere. I am thinking the operator description would be the most visible.
Some Bonsai.Vision operators that use OpenCV functions have constraints on the input
IplImagethat are not documented in the operator description or XML doc comments.For example, the
GoodFeaturesToTrackoperator only accepts an 8-bit or 32 bit, single channel image.The resulting error is hard to parse. For instance, using this operator on a normal RGB camera image results in:
It would be good to document these somewhere. I am thinking the operator description would be the most visible.