apollo::perception::lidar::PointPillars Class Reference

#include <point_pillars.h>

Collaboration diagram for apollo::perception::lidar::PointPillars:

Public Member Functions
	PointPillars (const bool reproduce_result_mode, const float score_threshold, const float nms_overlap_threshold, const std::string &pfe_torch_file, const std::string &scattered_torch_file, const std::string &backbone_torch_file, const std::string &fpn_torch_file, const std::string &bbox_head_torch_file)
	Constructor. More...
	~PointPillars ()
void	DoInference (const float *in_points_array, const int in_num_points, std::vector< float > *out_detections, std::vector< int > *out_labels)
	Call PointPillars for the inference. More...

Friends
class	TestClass

Constructor & Destructor Documentation

PointPillars()

apollo::perception::lidar::PointPillars::PointPillars	(	const bool	reproduce_result_mode,
		const float	score_threshold,
		const float	nms_overlap_threshold,
		const std::string &	pfe_torch_file,
		const std::string &	scattered_torch_file,
		const std::string &	backbone_torch_file,
		const std::string &	fpn_torch_file,
		const std::string &	bbox_head_torch_file
	)

Constructor.

Parameters

[in]	reproduce_result_mode	Boolean, if true, the output is reproducible for the same input
[in]	score_threshold	Score threshold for filtering output
[in]	nms_overlap_threshold	IOU threshold for NMS
[in]	pfe_torch_file	Pillar Feature Extractor Torch file path
[in]	scattered_torch_file	Pillar Feature scatter Torch file path
[in]	backbone_torch_file	Pillar Backbone Torch file path
[in]	fpn_torch_file	Pillar Fpn Torch file path
[in]	bbox_head_torch_file	Pillar Bbox Head Torch file path

Variables could be changed through point_pillars_detection

~PointPillars()

apollo::perception::lidar::PointPillars::~PointPillars

(

)

Member Function Documentation

DoInference()

void apollo::perception::lidar::PointPillars::DoInference	(	const float *	in_points_array,
		const int	in_num_points,
		std::vector< float > *	out_detections,
		std::vector< int > *	out_labels
	)

Call PointPillars for the inference.

Parameters

[in]	in_points_array	Point cloud array
[in]	in_num_points	Number of points
[out]	out_detections	Network output bounding box
[out]	out_labels	Network output object's label

This is an interface for the algorithm

Friends And Related Function Documentation

TestClass

friend class TestClass

friend

---

The documentation for this class was generated from the following file:

• modules/perception/lidar/lib/detector/point_pillars_detection/point_pillars.h