SegNet Jetson Inference with OpenCV


For segnet-camera.cpp in Jetson nano inference the software is capturing images from a gstCamera object. What I’m trying to do is to subscribe to an image_raw topic on ROS and use that image instead. Can anyone suggest a way to accomplish this? It looks like the first thing to do is to convert a cv::Mat object to a float pointer.



Here is an example in python interface.
Suppose you can apply the similar workflow in C++ version:


Hi @engietoday, please see the segnet node from ros_deep_learning, which uses jetson-inference: