Camera RAW Subband Compression for Low-Bandwidth Image Transmission
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing camera systems transmit RAW image data uncompressed, leading to increased power consumption and the need for expensive transmission cables, and existing compression methods do not efficiently handle pixel arrangements like Bayer and double density Bayer, resulting in decreased compression efficiency and aliasing noise.
Innovation Solution
The camera system performs subband division of image data with pixels of adjacent lines or columns as units, using wavelet or Haar transforms to efficiently compress and transmit data, allowing for real-time processing and reduced bandwidth requirements, and decompresses the data into original format for accurate signal processing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If RAW image data is transmitted in an uncompressed state, then transmission accuracy is maintained, but power consumption increases and transmission bandwidth requirements increase
Solution Approach 1:
The patent applies preliminary action by performing subband division and compression coding on RAW image data before transmission. The camera section compresses the data in advance using wavelet or Haar transforms, reducing the data volume that needs to be transmitted while maintaining essential image information. This preliminary compression reduces power consumption during transmission while preserving transmission accuracy through reversible transform coding.
Solution Approach 2:
The patent changes the parameter of data representation by transforming image data from spatial domain to frequency domain using wavelet or Haar transforms. This parameter change allows the data to be encoded more efficiently, reducing the number of bits required for transmission while maintaining image quality. The transform coefficients are then quantized and encoded, achieving compression without significant loss of transmission accuracy.
2Measurement precision
If RAW image data is transmitted in an uncompressed state, then image quality is preserved, but expensive transmission cables are required
Solution Approach 1:
The patent applies preliminary action by compressing RAW image data before transmission using subband division and transform coding. This pre-compression reduces the data volume significantly, allowing transmission over less expensive cables with lower bandwidth capacity. The compression is performed in advance at the camera section, so no expensive high-bandwidth cables are needed for transmitting uncompressed data.
Solution Approach 2:
The patent creates a compressed representation (copy) of the original RAW image data that contains the essential information in a reduced form. Instead of transmitting the full-resolution uncompressed data, a compressed version with transform coefficients is transmitted. This copied representation maintains image quality upon decompression while reducing transmission requirements and cable costs.
3Stability of the object's composition
If conventional compression methods are used on separated green images, then pixel position alignment is achieved, but compression efficiency decreases
Solution Approach 1:
The patent merges the previously separated G1 and G2 green image data back together before compression coding. Instead of compressing them as separate entities, the method combines them into a single green channel data stream, preserving the spatial correlations between adjacent pixels. This merging allows the compression algorithm to exploit the full spatial redundancy, significantly improving compression efficiency while maintaining proper pixel position alignment through the reversible transform process.
Solution Approach 2:
The patent treats all green pixel data (both G1 and G2) as homogeneous data belonging to the same color channel, rather than as separate entities. By applying the same compression processing uniformly to the combined green channel data, the method maintains consistency in pixel positioning while achieving higher compression ratios through efficient exploitation of spatial correlations across the entire green channel.
4Stability of the object's composition
If sub-sampling is performed to separate green images, then pixel position issues are resolved, but correlation between images is lost
Solution Approach 1:
The patent merges the separated G1 and G2 green image data back together before compression coding. Instead of compressing them as separate entities, the method combines them into a single green channel data stream, preserving the spatial correlations between adjacent pixels. This merging allows the compression algorithm to exploit the full spatial redundancy, significantly improving compression efficiency while maintaining proper pixel position alignment through the reversible transform process.
Solution Approach 2:
The patent uses subband division and transform coefficients as an intermediary representation that preserves correlations between green pixels without requiring separate image processing. The transform coefficients serve as a mediator that encodes the relationships between G1 and G2 pixels in a compressed form, allowing reconstruction of the correlated green channel data without losing the spatial relationships that would be lost in separate compression approaches.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach reduces power consumption and cable costs, enhances compression efficiency, and minimizes aliasing noise, enabling high-accuracy real-time processing of high-resolution images across various pixel arrangements.
Implementation Method 1
A camera section performs subband division of image data of a color whose pixel positions are alternately shifted from each other... The decompression processing section decompresses subband images with pixels of two upper and lower lines adjacent to each other or pixels of two left and right columns adjacent to each other as a unit, and synthesizes the subband images
Implementation Method 2
using wavelet or Haar transforms to efficiently compress and transmit data, allowing for real-time processing and reduced bandwidth requirements
Data Source
AI summary
Disclosed herein is a camera system including: a camera section including a subband dividing section configured to resolve image data of a color whose pixel positions are alternately shifted from each other into subband images, and a first transmission interface section configured to convert the subband images into a predetermined image signal, and output the image signal via a transmission line; and a camera control section including a second transmission interface section configured to convert the image signal input via the transmission line into the subband images, and an image decompressing section configured to decompress the subband images into the image data and synthesize the subband images into the image data output from the image pickup element.