Unlock instant, AI-driven research and patent intelligence for your innovation.

Monitoring camera demosaicing method based on calibration

A monitoring camera and demosaicing technology, applied in the field of image processing, can solve problems such as wrong interpolation direction, sensor interference by noise, deviation, etc., and achieve the effect of suppressing noise and improving interpolation accuracy

Active Publication Date: 2020-10-16
上海赛昉科技有限公司
View PDF8 Cites 2 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0003] Aiming at the deficiencies of the prior art, the present invention discloses a calibration-based monitoring camera demosaicing method, which solves the problem that in practical applications, taking the monitoring camera as an example, images under various illuminances need to be taken. When the illuminance is low, the sensor Seriously affected by noise, at this time the calculated color difference and gradient values ​​will deviate, resulting in the wrong direction of interpolation

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Monitoring camera demosaicing method based on calibration
  • Monitoring camera demosaicing method based on calibration
  • Monitoring camera demosaicing method based on calibration

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0036] This embodiment discloses as figure 1 A method for demosaicing a surveillance camera based on calibration shown, the method includes the following steps:

[0037] S1 calibrates the noise parameters of the sensor, adjusts the illuminance of the light source, records the magnification of the sensor, and calculates the statistical information of the 6 gray blocks at the bottom of the 24 color cards;

[0038] S2 changes the illuminance of the light source, repeats the process of S1 to calibrate the gradient and statistical information of the 6 gray blocks of the 24 color cards under different magnifications, and saves the magnification and the corresponding gray value and statistical information of each gray scale;

[0039] S3 finds the corresponding noise statistics in the calibration data based on the magnification of the sensor and the gray value of each pixel, subtracts a certain proportion of the noise gradient from the calculated gradient of the current point, and com...

Embodiment 2

[0059] In this embodiment, the method of demosaicing the monitoring camera based on calibration is combined with figure 2 To make a specific implementation operation, first calibrate the noise parameters of the sensor. In the light box environment, the object to be photographed is a 24-color card. Adjust the illumination of the light source so that the brightness of the image output by the sensor is suitable for human eyes. At this time, record the magnification of the sensor. Calculate the statistical information of the 6 gray blocks in the bottom row of the 24 color cards, such as gradient and variance, calculate the statistical information in the gray block point by point and calculate the average value in the entire gray block, and record the gray value of each gray block at the same time degree average. Gradient calculation can choose to refer to the sobel template to figure 2 Take the bayer image as an example, the formula for calculating the gradient in the horizonta...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention relates to the technical field of image processing, in particular to a monitoring camera demosaicing method based on calibration, which is characterized in that a sensor is set to have different gains, noise of data input into a demosaicing algorithm is calibrated, and calibration data is stored as prior information to be transmitted to the demosaicing algorithm. After input data isobtained through the demosaicing algorithm, the gradient of the data is calculated point by point, denoising processing is conducted on the gradient by referring to the calibrated noise parameters, and the interpolation direction of the current pixel point is judged according to the denoised gradient data; the method also includes judging whether the neighborhood of the current point is a flat area or a texture area according to noise parameters calibrated by the machine box, performing multi-directional interpolation on the flat area, and performing single-directional interpolation on pixel points of the texture area to remove noise without losing detail information.

Description

technical field [0001] The invention relates to the technical field of image processing, in particular to a method for demosaicing a monitoring camera based on calibration. Background technique [0002] The demosaic algorithm is an important module of the surveillance camera image processing system and the basis of the image effect. The current cameras or video cameras all use a single sensor. In order to obtain a color image, a color filter array (ColorFilter Array) is used to filter the incident light. Each pixel only accumulates energy for one color of red / green / blue light. In image processing, the process of restoring the other two colors based on the obtained data is demosaicing. General interpolation algorithms include bilinear interpolation and cubic interpolation. These algorithms have low computational complexity, but they do not treat all points in the image differently, resulting in serious false colors and zipper effects. At present, the interpolation algorithm...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
Patent Type & Authority Applications(China)
IPC IPC(8): G06T3/40G06T5/00G06T7/80
CPCG06T3/4015G06T3/4007G06T7/80G06T5/70
Inventor 何琦
Owner 上海赛昉科技有限公司