Vehicle vision system with video compression

• US 10,286,855 B2
• Filed: 03/22/2016
• Issued: 05/14/2019
• Est. Priority Date: 03/23/2015
• Status: Active Grant

• Current Assignee: Magna Electronics Incorporated (Magna International Incorporated)
• Original Assignee: Magna Electronics Incorporated (Magna International Incorporated)
• Inventors: Pflug, Goerg
• Primary Examiner(s): Werner, David N
• Application Number:
  US15/076,914
• Publication Number:
  US 20160286216A1
• Time in Patent Office: 1,148 Days
• CPC Class Codes: B60R 11/04 Mounting of cameras operative during drive; Arrangement of controls thereof relative to the vehicle
• CPC Class Codes: B60R 2300/102 using 360 degree surveillance camera system
• CPC Class Codes: B60R 2300/105 using multiple cameras
• CPC Class Codes: B60R 2300/30 characterised by the type of image processing
• CPC Class Codes: H04N 19/94 Vector quantisation
• CPC Class Codes: H04N 23/90 Arrangement of cameras or camera modules, e.g. multiple cameras in TV studios or sports stadiums
• CPC Class Codes: H04N 7/181 for receiving images from a plurality of remote sources

Abstract

A multi-camera vision system of a vehicle includes at least four cameras disposed at a vehicle and having respective fields of view exterior of the vehicle. Each of the cameras is operable to capture image data representative of the respective field of view. Captured image data is compressed at the respective camera and the compressed image data is communicated to a control unit. The control unit includes an image processor operable to process image data. The image processor processes image data frame by frame using an indirect context model and without time wise dependency.

Claims

1. A multi-camera vision system of a vehicle, said multi-camera vision system comprising:

at least four cameras disposed at a vehicle and having respective fields of view exterior of the vehicle;

wherein said at least four cameras comprises (i) at least one forward viewing camera having a field of view forward of the vehicle, (ii) at least one rearward viewing camera having a field of view rearward of the vehicle, (iii) at least one driver-side side viewing camera having a field of view sideward of the vehicle at the driver-side of the vehicle, and (iv) at least one passenger-side side viewing camera having a field of view sideward of the vehicle at the passenger-side of the vehicle;

wherein each of said cameras comprises a pixelated imaging array having a plurality of photosensing elements;

wherein each of said cameras is operable to capture image data within the respective field of view;

wherein captured image data is compressed at the respective camera and wherein the compressed image data is communicated to a control unit;

wherein said control unit comprises an image processor;

wherein said image processor processes image data frame by frame using an indirect context model and without time wise dependency;

wherein compressed image data is received at said control unit and decompressed before processing of image data by said image processor;

wherein captured image data is compressed via a lossy video data compression and compressed image data is decompressed at said control unit via a decompression codec based on vector quantization using said indirect context model to select from available vectors; and

wherein captured image data is compressed at a compression rate that depends on a selected vector length and color space.

2. The multi-camera vision system of claim 1, wherein, after a full frame is coded, a vector table becomes updated.

3. The multi-camera vision system of claim 2, wherein a sum of vector table elements is divided by a counter element.

4. The multi-camera vision system of claim 1, wherein a pixel or sequence of consecutive pixels are coded as a decision bit by deciding a vector duplet indexed by a context.

5. The multi-camera vision system of claim 4, wherein the vector duplet indexed by the context is added to a predicted value of the pixel or sequence of consecutive pixels to determine an error.

6. The multi-camera vision system of claim 5, wherein the error is coded and entered in a bit field for the according source pixel'"'"'s position.

7. The multi-camera vision system of claim 1, wherein a context is based on the context of prior decision bits surrounding a pixel.

8. The multi-camera vision system of claim 1, wherein a data channel bandwidth for captured image data is constant.

9. The multi-camera vision system of claim 1, wherein compressed image data is communicated to said control unit via an Ethernet communication.

10. The multi-camera vision system of claim 1, wherein a pixel is processed by the context model based on at least one other pixel adjacent to the pixel being processed.

11. A multi-camera vision system of a vehicle, said multi-camera vision system comprising:

at least four cameras disposed at a vehicle and having respective fields of view exterior of the vehicle;

wherein said at least four cameras comprises (i) at least one forward viewing camera having a field of view forward of the vehicle, (ii) at least one rearward viewing camera having a field of view rearward of the vehicle, (iii) at least one driver-side side viewing camera having a field of view sideward of the vehicle at the driver-side of the vehicle, and (iv) at least one passenger-side side viewing camera having a field of view sideward of the vehicle at the passenger-side of the vehicle;

wherein each of said cameras comprises a pixelated imaging array having a plurality of photosensing elements;

wherein each of said cameras is operable to capture image data within the respective field of view;

wherein captured image data is compressed at the respective camera and wherein the compressed image data is communicated to a control unit;

wherein compressed image data is communicated to said control unit via an Ethernet communication;

wherein said control unit comprises an image processor;

wherein compressed image data is received at said control unit and decompressed before processing of image data by said image processor;

wherein said image processor processes image data frame by frame using an indirect context model and without time wise dependency; and

wherein captured image data is compressed via a lossy video data compression and compressed image data is decompressed at said control unit via a decompression codec based on vector quantization using said indirect context model to select from available vectors.

12. The multi-camera vision system of claim 11, wherein captured image data is compressed at a compression rate that depends on a selected vector length and color space, and wherein, after a full frame is coded, a vector table becomes updated.

13. The multi-camera vision system of claim 11, wherein a pixel or sequence of consecutive pixels are coded as a decision bit by deciding a vector duplet indexed by a context, and wherein the vector duplet indexed by the context is added to a predicted value of the pixel or sequence of consecutive pixels to determine an error, and wherein the error is coded and entered in a bit field for the according source pixel'"'"'s position.

14. The multi-camera vision system of claim 11, wherein a pixel is processed by the context model based on at least one other pixel adjacent to the pixel being processed.

15. A multi-camera vision system of a vehicle, said multi-camera vision system comprising:

at least four cameras disposed at a vehicle and having respective fields of view exterior of the vehicle;

wherein said at least four cameras comprises (i) at least one forward viewing camera having a field of view forward of the vehicle, (ii) at least one rearward viewing camera having a field of view rearward of the vehicle, (iii) at least one driver-side side viewing camera having a field of view sideward of the vehicle at the driver-side of the vehicle, and (iv) at least one passenger-side side viewing camera having a field of view sideward of the vehicle at the passenger-side of the vehicle;

wherein each of said cameras comprises a pixelated imaging array having a plurality of photosensing elements;

wherein each of said cameras is operable to capture image data within the respective field of view;

wherein captured image data is compressed at the respective camera and wherein the compressed image data is communicated to a control unit;

wherein said control unit comprises an image processor;

wherein compressed image data is received at said control unit and decompressed before processing of image data by said image processor;

wherein said image processor processes image data frame by frame using an indirect context model and without time wise dependency;

wherein a pixel or sequence of consecutive pixels are coded as a decision bit by deciding a vector duplet indexed by a context; and

wherein captured image data is compressed via a lossy video data compression and compressed image data is decompressed at said control unit via a decompression codec based on vector quantization using said indirect context model to select from available vectors, and wherein captured image data is compressed at a compression rate that depends on a selected vector length and color space.

16. The multi-camera vision system of claim 15, wherein, after a full frame is coded, a vector table becomes updated, and wherein the vector duplet indexed by the context is added to a predicted value of the pixel or sequence of consecutive pixels to determine an error, and wherein the error is coded and entered in a bit field for the according source pixel'"'"'s position.