I am looking for a pre-trained/pre-made .hef file that will specifically ONLY detect airplanes from satellite imagery. I am wondering if I can just download it from somewhere. I have a .tflite model that is converts to .onnx but, doest not convert .hef.
That is very specific. I do not believe we have one of these ready out-of-the-box.
You can convert TFLite to HEF using the Hailo Dataflow Compiler.
If you can provide some more detailed error messages we can help you convert the model.
Your best option is to use models that are already in the Hailo Model Zoo especially the YOLO family. That will simplify the conversion process.
Have you worked trough the tutorials in the Hailo AI Software Suite Docker? The tutorials guide you step-by-step through the workflow. You can start the Jupyter notebook server with the tutorial notebooks by calling the following command inside the Docker.
hailo tutorial
So, I was able to convert TFLite to HEF. Now, it looks like when the object is detected, I get this error message:
~/hailo-apps/hailo_apps/python/standalone_apps/object_detection$ ./object_detection.py -n rareplanesv3.hef -i test10.png
INFO | common.core | Using HEF from path: /home/gts/hailo-apps/hailo_apps/python/standalone_apps/object_detection/rareplanesv3.hef
Exception in thread Thread-2 (visualize):
Traceback (most recent call last):
File “/usr/lib/python3.10/threading.py”, line 1016, in _bootstrap_inner
self.run()
File “/usr/lib/python3.10/threading.py”, line 953, in run
self._target(*self._args, **self._kwargs)
File “/home/gts/hailo-apps/hailo_apps/python/core/common/toolbox.py”, line 861, in visualize
frame_with_detections = callback(original_frame, inference_result)
File “/home/gts/hailo-apps/hailo_apps/python/standalone_apps/object_detection/object_detection_post_process.py”, line 38, in inference_result_handler
detections = extract_detections(original_frame, infer_results, config_data) # Should return dict with boxes, classes, scores
File “/home/gts/hailo-apps/hailo_apps/python/standalone_apps/object_detection/object_detection_post_process.py”, line 140, in extract_detections
bbox, score = det[:4], det[4]
IndexError: invalid index to scalar variable.
SUCCESS | main | Inference was successful!
INFO | main | Results have been saved in /home/gts/hailo-apps/hailo_apps/python/standalone_apps/object_detection/output