Vehicle navigation system with vision system preprocessor using MPEG encoder
First Claim
1. A vehicle navigation system, comprising:
- at least one sensor for capturing image frames, each image frame containing pixel color data;
a pre-processor coupled to said at least one camera for receiving image frames, wherein the pre-processor converts the pixel color data into luminance and chrominance data; and
a processor that processes the luminance and chrominance data to identify at least one of an object and a terrain characteristic, wherein the processor transforms the luminance data into frequency components, and wherein the frequency components from the luminance component are two-dimensional frequency components having a high spatial frequency and a low spatial frequency, and wherein the processor calculates the ratio between the high and low spatial frequencies to determine a slope of the terrain relative to the vehicle'"'"'s line of sight.
5 Assignments
0 Petitions
Accused Products
Abstract
A navigation system for an unmanned ground vehicle (UGV) includes an MPEG encoder that pre-processes frame data from a plurality of cameras on the UGV before the data is used to generate maps that identify objects in the vehicle'"'"'s field of view and that identify terrain characteristics. The MPEG encoder converts the red-green-blue (RGB) data into luminance-chrominance (YUV) data. The luminance and chrominance data for each frame are divided into blocks for determining motion vectors, average chrominance values, and spatial frequencies for each block. A range map identifying objects is generated from the motion vectors and average chrominance values, and a slope map identifying terrain characteristics is generated from the spatial frequencies.
108 Citations
17 Claims
-
1. A vehicle navigation system, comprising:
-
at least one sensor for capturing image frames, each image frame containing pixel color data;
a pre-processor coupled to said at least one camera for receiving image frames, wherein the pre-processor converts the pixel color data into luminance and chrominance data; and
a processor that processes the luminance and chrominance data to identify at least one of an object and a terrain characteristic, wherein the processor transforms the luminance data into frequency components, and wherein the frequency components from the luminance component are two-dimensional frequency components having a high spatial frequency and a low spatial frequency, and wherein the processor calculates the ratio between the high and low spatial frequencies to determine a slope of the terrain relative to the vehicle'"'"'s line of sight. - View Dependent Claims (2)
-
-
3. A vehicle navigation system, comprising:
-
at least one sensor for capturing image frames, each image frame containing pixel color data;
a pre-processor coupled to said at least one camera for receiving image frames, wherein the pre-processor converts the pixel color data into luminance and chrominance data; and
a processor that processes the luminance and chrominance data to identify at least one of an object and a terrain characteristic, wherein the processor generates a plurality of macroblocks for each image frame and transforms the chrominance data of each macroblock to obtain an average chrominance value, generates a motion vector of each macroblock from the luminance data, and wherein the processor includes;
a range calculator that identifies the range corresponding to each macro-block from the motion vector;
a quantizer that quantizes the chrominance value for each macro-block to obtain an average block color; and
a segmenter that segments the frame by merging adjacent macro-blocks having substantially the same range and substantially the same block color to identify the object.
-
-
4. A vehicle navigation system, comprising:
-
a plurality of sensors for capturing image frames, each image frame containing pixel color data;
a pre-processor coupled to said plurality of sensors for receiving said image frames, wherein the pre-processor converts the pixel color data into luminance and chrominance data; and
a processor that processes the luminance and chrominance data to generate at least one of a range map that identifies an object and object range and a slope map that identifies a terrain characteristic, wherein the processor transforms the luminance and chrominance data and generates motion vectors and image segments for generating said range map and slope map. - View Dependent Claims (5, 6, 7, 8, 9, 10, 11)
a range calculator that identifies the range corresponding to each macroblock from the motion vector;
a quantizer that quantizes the chrominance value for each macroblock to obtain an average block color; and
a segmenter that segments the frame by merging adjacent macroblocks having substantially the same range and substantially the same block color to identify the object.
-
-
12. A method for navigating a vehicle, comprising the steps of:
-
obtaining a first image frame and a second image frame, wherein the image frames contain pixel color data;
converting the pixel color data into luminance data and chrominance data;
detecting at least one of an object and object range using the chrominance data; and
detecting a terrain characteristic using the luminance data;
creating a range map of the object identified in the object detecting step; and
creating a slope map from the terrain characteristics identified in the terrain characteristic detecting step, wherein the range map and the slope map are used to navigate the vehicle. - View Dependent Claims (13, 14, 15, 16, 17)
dividing each frame into a plurality of macroblocks;
deriving a motion vector for each macroblock from the luminance value, the motion vector indicating a range;
obtaining an average block color for each macroblock; and
merging adjacent macroblocks having substantially the same average block color and substantially the same range to generate the range map identifying the presence and range of objects.
-
-
14. The method of claim 13, wherein the step of obtaining the average block color includes the step of transforming the chrominance data to obtain an average chrominance value used to generate the average block color.
-
15. The method of claim 14, wherein the transforming step is conducted via discrete cosine transformation.
-
16. The method of claim 12, wherein the step of creating a slope map includes the steps of:
-
dividing the luminance data for each frame into blocks;
transforming the luminance data to obtain two-dimensional frequency components having a high spatial frequency and a low spatial frequency for each block;
calculating a ratio between the high and low spatial frequencies to determine a slope of the terrain relative to the vehicle'"'"'s line of sight; and
calculating an angle between the low spatial frequency and a gravity vector to determine a lateral slope of the terrain.
-
-
17. The method of claim 16, wherein the transforming step is conducted via discrete cosine transformation.
Specification