Long Range Object Detector #16
Conversation
* ZED Naming conflicts * OD in the sim and on ZED * style * style
| std::vector<int64_t> outputTensorSize; | ||
|
|
||
| // Additional space for params that one algorithm might need and not others | ||
| std::vector<float> buffer; |
There was a problem hiding this comment.
rethink the buffer, unspecified size is not ideal
There was a problem hiding this comment.
Yes, this was a bad feature making non static calls is def the move
| } | ||
|
|
||
| int rows = output.cols; | ||
| int dimensions = output.rows; |
There was a problem hiding this comment.
can you name this more verbosely
| @@ -50,18 +70,43 @@ namespace mrover { | |||
|
|
|||
There was a problem hiding this comment.
image should be const ref
There was a problem hiding this comment.
Ok, I changed publishDetectedObjects to const ref, drawDetectionBoxes to just ref and rgbImage in preprocessYOLOv8Input to const ref, lmk if i missed any
| public: | ||
| explicit ImageObjectDetector(); | ||
|
|
||
| auto getTagBearing(cv::InputArray image, cv::Rect const& box) const -> float; |
There was a problem hiding this comment.
don't include in this PR
| } | ||
|
|
||
| // Be verbose about the input tensor size | ||
| auto outputTensorSize = getInputTensorSize(); |
There was a problem hiding this comment.
auto outputTensorSize = getOutputTensorSize();
| std::ofstream outputFileStream{enginePath, std::ios::binary}; | ||
| outputFileStream.write(static_cast<char const*>(trtModelStream->data()), trtModelStream->size()); | ||
| std::ofstream outputFileStream{mEngineModelPath, std::ios::binary}; | ||
| outputFileStream.write(static_cast<char const*>(trtModelStream->data()), static_cast<int32_t>(trtModelStream->size())); |
There was a problem hiding this comment.
narrowing conversion, but I made it more clear
| @@ -153,7 +182,8 @@ auto Inference::prepTensors() -> void { | |||
|
|
|||
| assert(mContext); | |||
| // Create an appropriately sized output tensor | |||
| auto const [nbDims, d] = mEngine->getTensorShape(OUTPUT_BINDING_NAME); | |||
| // TODO: Fix this | |||
| auto const [nbDims, d] = mEngine->getTensorShape(mOutputTensorName.c_str()); | |||
There was a problem hiding this comment.
this TODO still relevant?
| @@ -170,3 +200,28 @@ auto Inference::getBindingInputIndex(IExecutionContext const* context) -> int { | |||
| // Returns the id for the input tensor | |||
| return context->getEngine().getTensorIOMode(context->getEngine().getIOTensorName(0)) != TensorIOMode::kINPUT; // 0 (false) if bindingIsInput(0), 1 (true) otherwise | |||
| } | |||
|
|
|||
|
|
|||
| auto Inference::getInputTensorSize() -> std::vector<int64_t>{ | |||
There was a problem hiding this comment.
Do you really need two separate functions for input and output?
| // Fill in the output Detections Vector | ||
| for (int i: nmsResult) { | ||
| Detection result; | ||
|
|
||
| //Fill in the id and confidence for the class | ||
| result.classId = classIds[i]; | ||
| result.confidence = confidences[i]; | ||
|
|
||
| //Fill in the class name and box information | ||
| result.className = model.classes[result.classId]; | ||
| result.box = boxes[i]; | ||
|
|
||
| //Push back the detection into the for storagevector | ||
| detections.push_back(result); | ||
| } |
There was a problem hiding this comment.
Can this be made more concise?
| for (auto const& [classId, className, confidence, box]: detections) { | ||
| mrover::msg::ImageTarget target; | ||
| target.name = className; | ||
| target.bearing = getTagBearing(blobSizedImage, box); | ||
| targets.targets.push_back(target); | ||
| } |
There was a problem hiding this comment.
surely this could be emplace_back
| @@ -112,7 +112,7 @@ def paintEvent(self, event): | |||
| self.resize(self.renderer.defaultSize()) | |||
| self.renderer.render(painter) | |||
|
|
|||
| def update(self): | |||
| def update(self): # type: ignore[override] | |||
There was a problem hiding this comment.
I feel this is not relevant to the PR
Summary
Closes #14
What features did you add, bugs did you fix, etc?
I added the long range object detector
Did you add documentation to the wiki?
No, its already on the wiki
How was this code tested?
Tested on usb camera while also running the regular object detector
Did you test this in sim?
No
Did you test this on the rover?
No
Did you add unit tests?
No