apollo::perception::lidar::ScatterCuda 的协作图:

Public 成员函数
	ScatterCuda (const int num_threads, const int grid_x_size, const int grid_y_size)
	Constructor

void	DoScatterCuda (const int pillar_count, int x_coors, int y_coors, float pfe_output, float scattered_feature)
	Call scatter cuda kernel

详细描述

在文件 scatter_cuda.h 第 46 行定义.

构造及析构函数说明

apollo::perception::lidar::ScatterCuda::ScatterCuda	(	const int	num_threads,
		const int	grid_x_size,
		const int	grid_y_size
	)

Constructor

参数

[in]	num_threads	The number of threads to launch cuda kernel
[in]	grid_x_size	Number of pillars in x-coordinate
[in]	grid_y_size	Number of pillars in y-coordinate

Captital variables never change after the compile

void apollo::perception::lidar::ScatterCuda::DoScatterCuda	(	const int	pillar_count,
		int *	x_coors,
		int *	y_coors,
		float *	pfe_output,
		float *	scattered_feature
	)

Call scatter cuda kernel

参数

[in]	pillar_count	The valid number of pillars
[in]	x_coors	X-coordinate indexes for corresponding pillars
[in]	y_coors	Y-coordinate indexes for corresponding pillars
[in]	pfe_output	Output from Pillar Feature Extractor
[out]	scattered_feature	Gridmap representation for pillars' feature

Allocate pillars in gridmap based on index(coordinates) information

该类的文档由以下文件生成:

modules/perception/lidar_detection/detector/point_pillars_detection/scatter_cuda.h