Online Lidar intensity normalization
Aspects of the present disclosure involve a vehicle computer system comprising a computer-readable storage medium storing a set of instructions, and a method for online light detection and ranging (Lidar) intensity normalization. Consistent with some embodiments, the method may include accumulating point data output by a channel of a Lidar unit during operation of an autonomous or semi-autonomous vehicle. The accumulated point data includes raw intensity values that correspond to a particular surface type. The method further includes calculating a median intensity value based on the raw intensity values and generating an intensity normalization multiplier for the channel based on the median intensity value. The intensity normalization multiplier, when applied to the median intensity value, results in a reflectivity value that corresponds to the particular surface type. The method further includes applying the intensity normalization multiplier to the point data output by the channel to produce normalized intensity values.
1 . A vehicle computing system comprising:
one or more processors; and
a non-transitory computer-readable medium storing instructions that are executable by the one or more processors to cause the vehicle computing system to perform operations comprising:
obtaining point data output by a first channel from among a plurality of channels associated with a sensor device of a vehicle, the point data comprising a plurality of raw intensity values that correspond to a particular surface type;
calculating a particular intensity value based on the plurality of raw intensity values;
generating an intensity normalization multiplier for the first channel based on the particular intensity value;
based on the intensity normalization multiplier, determining a normalized intensity value for subsequent point data output by the first channel;
recalculating the intensity normalization multiplier for the first channel at a subsequent time interval to account for a change in a temperature of an environment of the sensor device;
determining a motion path for the vehicle based on the normalized intensity value; and
controlling operations of the vehicle based on the motion path.
2 . The vehicle computing system of claim 1 , wherein application of the intensity normalization multiplier to the particular intensity value results in a reflectivity value that corresponds to the particular surface type.
3 . The vehicle computing system of claim 1 , wherein the operations further comprise:
obtaining additional point data output by a second channel from among the plurality of channels of the sensor device;
generating an additional intensity normalization multiplier for the second channel; and
applying the additional intensity normalization multiplier to the additional point data output by the second channel.
4 . The vehicle computing system of claim 1 , wherein the operations further comprise:
updating the intensity normalization multiplier based on the subsequent point data output by the first channel; and
applying the updated intensity normalization multiplier to further point data output by the first channel.
5 . The vehicle computing system of claim 1 , wherein the sensor device is a light detection and ranging (Lidar) system or a Radar system.
6 . The vehicle computing system of claim 1 , wherein the vehicle is an autonomous vehicle.
7 . The vehicle computing system of claim 1 , wherein the vehicle is a truck.
8 . An autonomous vehicle comprising:
at least one least one Lidar sensor unit configured to:
obtain point data output by a first channel from among a plurality of channels of the Lidar sensor unit, the point data comprising a plurality of respective raw intensity values for a plurality of data points; and
one or more processors configured to:
determine a reflectivity value for an expected surface type for at least a subset of data points of the plurality of data points based on the raw intensity values for the subset of data points;
generate normalized point data for the first channel of the plurality of channels of the Lidar sensor unit based on the reflectivity value for the expected surface type for the subset of data points; and
perform one or more operations for the autonomous vehicle based on the normalized point data;
wherein generating normalized point data for the first channel of the Lidar sensor unit comprises:
calculating a particular intensity value based on the plurality of raw intensity values;
generating an intensity normalization multiplier for the first channel based on the particular intensity value;
determining a normalized intensity value for the normalized point data based on the intensity normalization multiplier, wherein the normalized point data comprises subsequent point data output by the first channel; and
recalculating the intensity normalization multiplier for the first channel at a subsequent time interval to account for a change in a temperature of an environment of the Lidar sensor unit.
9 . The vehicle computing system of claim 1 , wherein the operations further comprise:
recalculating the intensity normalization multiplier for the first channel to account for a change in alignment of an emitter and a detector in the first channel.
10 . A light detection and ranging (Lidar) system for a vehicle, the Lidar system comprising:
at least one Lidar sensor unit configured to:
obtain point data output by a first channel from among a plurality of channels of the Lidar sensor unit, the point data comprising a plurality of respective raw intensity values for a plurality of data points associated with the first channel; and
one or more processors configured to:
determine a reflectivity value for an expected surface type for at least a subset of data points of the plurality of data points based on the raw intensity values for the subset of data points;
generate normalized point data for the first channel of the Lidar sensor unit based on the reflectivity value for the expected surface type for the subset of data points; and
provide the normalized point data to a computing system of an autonomous vehicle;
wherein generating normalized point data for the first channel of the Lidar sensor unit comprises:
calculating a particular intensity value based on the plurality of raw intensity values;
generating an intensity normalization multiplier for the first channel based on the particular intensity value;
determining a normalized intensity value for the normalized point data based on the intensity normalization multiplier, wherein the normalized point data comprises subsequent point data output by the first channel; and
recalculating the intensity normalization multiplier for the first channel at a subsequent time interval to account for a change in a temperature of an environment of the Lidar sensor unit.
11 . The Lidar system of claim 10 , wherein determining the reflectivity value for the expected surface type for at least the subset of data points of the plurality of data points comprises:
accessing a reference data structure comprising one or more historic data points corresponding to the expected surface type; and
determining the expected surface type based on the reference data structure.
12 . The Lidar system of claim 10 , wherein the reflectivity value is a ground truth reflectivity value for the expected surface type, the ground truth reflectivity value comprising a known measure of a reflective characteristic of the expected surface type.
13 . The Lidar system of claim 10 , wherein generating normalized point data for the first channel of the Lidar sensor unit comprises:
recalculating the intensity normalization multiplier for the first channel to account for a change in alignment of an emitter and a detector in the first channel.
14 . The Lidar system of claim 10 , wherein the at least one Lidar sensor unit is further configured to obtain point data output by a second channel from among the plurality of channels of the Lidar sensor unit, the point data output by the second channel comprising a plurality of data points associated with the second channel, and
wherein the one or more processors are further configured to generate normalized point data for the second channel of the Lidar sensor unit based on a reflectivity value for an expected surface type for at least a subset of data points associated with the second channel.
15 . The autonomous vehicle of claim 8 , wherein the one or more operations for the autonomous vehicle comprise predicting a future location of an object within an environment of the autonomous vehicle.
16 . The autonomous vehicle of claim 8 , wherein generating normalized point data for the first channel of the plurality channels of the Lidar sensor unit comprises:
recalculating the intensity normalization multiplier for the first channel to account for a change in alignment of an emitter and a detector in the first channel.
17 . The autonomous vehicle of claim 8 , wherein the one or more operations for the autonomous vehicle comprise determining a motion path for the autonomous vehicle.
18 . The autonomous vehicle of claim 17 , wherein the one or more operations for the autonomous vehicle further comprise controlling the autonomous vehicle based on the motion path.
19 . The autonomous vehicle of claim 8 , wherein the one or more operations for the autonomous vehicle comprise localizing the autonomous vehicle within an environment of the autonomous vehicle.
20 . The autonomous vehicle of claim 8 , wherein the one or more operations for the autonomous vehicle comprise identifying an object within an environment of the autonomous vehicle.