Hybrid-view LIDAR-based object detection

An autonomous vehicle computing system, comprising:

• a LIDAR system configured to transmit ranging signals relative to an autonomous vehicle and to generate LIDAR data, wherein the LIDAR data includes a plurality of LIDAR points detected by the LIDAR system;
• one or more processors;
• a classification model, wherein the classification model has been trained to classify cells of LIDAR data;
• and
• at least one tangible, non-transitory computer readable medium that stores instructions that, when executed by the one or more processors, cause the one or more processors to perform operations, the operations comprising:
• accessing a predetermined set of objects of interest;
• determining one or more cell statistics characterizing the LIDAR data corresponding to the cells, the one or more cell statistics characterizing respective LIDAR data corresponding to a respective cell and comprising one or more parameters associated with a distribution of a plurality of detected LIDAR data points within the respective cell, a power of the plurality of detected LIDAR data points, or an intensity of the plurality of detected LIDAR data points projected onto the respective cell;
• providing the one or more cell statistics as input to the classification model;
• receiving, as output of the classification model, a classification indicating whether for each cell includes a detected object from the predetermined set of objects of interest;
• and
• controlling one or more vehicle controls of the autonomous vehicle based, at least in part, on a determination whether the cell includes at least one object included in the predetermined set of objects of interest.

An autonomous vehicle, comprising:

• a sensor system comprising at least one LIDAR system configured to transmit ranging signals relative to the autonomous vehicle and to generate LIDAR data;
• and
• a vehicle computing system comprising:
• one or more processors;
• and
• at least one tangible, non-transitory computer readable medium that stores instructions that, when executed by the one or more processors, cause the one or more processors to perform operations, the operations comprising:
• receiving LIDAR data from the sensor system, wherein the LIDAR data includes a plurality of LIDAR points detected by the sensor system;
• generating a top-view representation of the LIDAR data that is discretized into a grid of multiple cells, each cell representing a column in three-dimensional space;
• determining one or more cell statistics characterizing the LIDAR data corresponding to the cells, the one or more cell statistics characterizing respective LIDAR data corresponding to a respective cell and comprising one or more parameters associated with a distribution of a plurality of detected LIDAR data points within the respective cell, a power of the plurality of detected LIDAR data points, or an intensity of the plurality of detected LIDAR data points projected onto the respective cell;
• determining a feature extraction vector for each cell by aggregating one or more cell statistics of surrounding cells at one or more different scales;
• determining a classification for each cell based at least in part on the feature extraction vector for each cell;
• and
• controlling, by the vehicle computing system, one or more vehicle controls of the autonomous vehicle based, at least in part, on the classification for each cell.