Table of Contents
- Overview of Visualizer Settings
- Overview of Perception Setting
The "Settings" tab of the main sidebar is home to all the visualizer settings. There are a few settings categories:
Sensor Reading Settings:
The Sensor Reading Settings section provides settings that augment the appearance of the visualized sensor readings. Some settings are intuitive (such as radius) while others require a little explanation.
|Radius||Changes the size of the sensor reading data points|
|Color Scheme||Changes the color scheme used to color the sensor reading data points. Color schemes can be uniform or attribute dependent. For example, the color scheme "Gradient on Radial Distance" will color the data points based on their distance from the origin.|
|Color Map||The color map used to color sensor readings on a gradient.|
|Legend Saturation Limits||Adjusts the limits of the color map gradient such that the gradient is completely saturated at the limits. Suppose you set the "Color Scheme" to be a gradient on signal strength, and that you set the "Color Map" to be grayscale. Then adjusting the saturation limits to something like 100 and 200, would cause any sensor reading with a signal strength below 100 to be colored black and any reading with a signal strength above 200 to be colored white. In the range 100 to 200 the sensor readings would be colored on a gradient between black and white.|
|Sweep Decay History||Settings for visualizing previous sweeps as an overlaid history. The toggle switch will turn on the decay sweeps and the count slider will change how many previous sweeps are drawn. Note that this will increase the processing requirements of the visualizer. The translucency settings changes the presence or type of translucency scheme to better differentiate previous sweeps.|
The background settings provide a means to adjust the aesthetics of the visualizer, including:
|Background Color||Changes the color of the visualizer background.|
|Grid Color||Changes the color of the visualizer grid.|
|Grid Bounds||Changes the maximum bounds of the grid (ie: 1500 means a cartesian grid extending from -1500:1500 in both x and y, and a polar grid extending to a radius of 1500).|
|Cartesian Grid||The toggle switch toggles the presence of the cartesian grid. The input box permits the user to change the size of a cartesian grid square.|
|Polar Grid||The toggle switch toggles the presence of the polar grid. The input box permits the user to change the distance between radial delineations on the polar grid.|
Eventually the SDK will include numerous methods/solutions to common perception tasks. However these features were put on hold until engineering resources free up to incorporate them in future releases of the software. Currently the only included features are basic boundary detection via line-extraction, and basic Cartesian clustering.
Boundary Detection Settings
The boundary detection settings provide a means to tune the line-extraction algorithm (split-merge).
|Toggle Detection||Toggles the boundary detection on or off.|
|Min Radius||Defines a minimum radius for points to be considered in the line extraction. This prevents lines from forming on sensor readings near the sensor.|
|Distance Threshold||The max distance a point can be from a line to be considered a part of that line.|
|Collinearity Threshold||A metric for how collinear two lines segments have to be before they are merged into a single segment. Lower values are more strict, that is the segments must be more collinear in order to be merged. Higher values will attempt to merge lines that aren't as collinear.|
|Min Num Pts in Line||The minimum number of points for a line to remain. Lines made up of fewer points will be removed. This is the only currently implemented post-processing technique.|
|Use Intermediate Line Fitting||Whether or not to use intermediate line fitting during the split portion of the split-merge line extraction algorithm. If true, the recursive split step will consider the most distant point to the line that is formed as a fit to a subset of points. If false, the recursive split step will consider the most distant point to the line that runs between the two most distant points in a subset of points.|
The clustering settings provide a means to tune the basic clustering methods. Currently only common Cartesian based clustering algorithms are provided, but soon support will be added for polar based clustering that is more performant with 2D scanning LiDAR.
|Toggle Clustering||Toggles clustering on or off.|
|Clustering Method||Selects a method for clustering.|
|Neighborhood Radius||The size of the neighborhood radius in cm.|
|Minimum Cluster Size||The number of points in a radius to form a cluster.|
|Number of Clusters||The desired number of clusters (for k-means only)|