Grid Detect

GridDetect processes high-resolution 3D point cloud data to deliver real-time Bird's Eye View (BEV) occupancy maps and object tracking. It uses a deterministic signal processing approach that does not require training. It supports two detection modes that can be selected via the enable_grid_detect parameter in the master config.

Modes

Value	Mode	Description
0	Disabled	Grid Detect is off. No obstacle data is published.
1	Global	Detects obstacles over a wide area. Best suited for flat surfaces at longer ranges.
2	Local	Detects obstacles based on local variations in the point cloud.

Parameters

The following parameters are available in the master config:

Parameter	Type	Default	Description
enable_grid_detect	int	0	Grid Detect mode (0 = disabled, 1 = Global, 2 = Local).
griddetect_sensitivity	float	0.9	Detection sensitivity (0-1). Lower values mean less sensitive (fewer detections). Higher values mean more sensitive (more detections).
obstacle_x_min	float	-15	Minimum x-coordinate of the detection volume in meters.
obstacle_x_max	float	15	Maximum x-coordinate of the detection volume in meters.
obstacle_y_min	float	-3	Minimum y-coordinate of the detection volume in meters.
obstacle_y_max	float	3	Maximum y-coordinate of the detection volume in meters.
obstacle_z_min	float	0	Minimum z-coordinate (depth) of the detection volume in meters.
obstacle_z_max	float	100	Maximum z-coordinate (depth) of the detection volume in meters.