CN102377953A - Filtering and denoising system and filtering and denoising method for video data - Google Patents

Abstract

The invention relates to a filtering and denoising system and a filtering and denoising method for video data. An interframe data processing module, a frame data caching module and a frame data output control module are connected to form a data path, subtraction is carried out between the previous frame data and the current frame, a result worked out by the operation is multiplied by a filtering coefficient and then added by the current frame, so that the filtered current frame data are worked out, the frame data caching module comprises a DDR2 (Double Data Rate 2) memory, and has two caches, each cache caches the data of a frame of video image, so that the invention can be adapted to the denoising of the video images of an object moving at low speed, the time-weighted average of the image signals is worked out according to the frame period, so that the image signals are almost not affected, consequently, not only can the edges in the images be kept well, but also noise can be greatly reduced, and thereby the purpose of improving the signal-to-noise ratio of the images is achieved.

Description

The filtering noise reduction system of video data and filtering noise-reduction method

Technical field

The present invention relates to the technical field of video monitoring; Be a kind ofly to form specifically by interframe data processing module, frame data cache module and frame data output control module; Be applied to the noise reduction of real-time video, effectively the filtering noise reduction system and the filtering noise-reduction method of the video data of filtering random noise disturbance.

Background technology

The main target of video image denoising is a filtering random noise wherein, has kept detailed information simultaneously as far as possible, and requires to reduce in the video image behind noise reduction because the vision that filtering is introduced degrades.

The noise of video image denoising technology in not only can the filtering video image, improve the video image subjective visual quality do, and significant for subsequent treatment tasks such as compressed encoding, target recognition and tracking, frame frequency liftings.Existing video image denoising algorithm can be divided into two types: early stage pixel domain noise reduction algorithm and conversion territory noise reduction algorithm in recent years.

According to the filter range of filter, filtering algorithm when the pixel domain noise reduction algorithm can be divided into the time-domain filtering algorithm with sky.The time-domain filtering algorithm utilizes correlation on the video image time domain to suppress noise, obtains time domain prediction based on the motion estimation/motion compensation method usually; Filtering algorithm then is a correlation filtering noise when utilizing empty in the video image three dimensions when empty.The major defect of pixel domain noise reduction algorithm is exactly to introduce easily in the video image behind noise reduction that time domain degrades, spatial domain such as level and smooth excessively degrades; Up to the present also do not have a kind of other noise reduction algorithm of multiple noise level that is fit to, adopt adaptation mechanism can improve the defective of pixel domain noise reduction algorithm.In addition, the estimation that adopt are obtained the relevant information on the time domain more in the pixel domain noise reduction algorithm, but the existence of noise influences the accuracy of estimation easily, thereby reduce anti-acoustic capability.

Summary of the invention

The technical problem that the present invention will solve provides a kind ofly to be made up of interframe data processing module, frame data cache module and frame data output control module; Be applied to the noise reduction of real-time video, effectively the filtering noise reduction system and the filtering noise-reduction method of the video data of filtering random noise disturbance.

The technical scheme that the present invention takes for the technical problem that exists in the solution known technology is:

The filtering noise reduction system of video data of the present invention; Comprise interframe data processing module, frame data cache module and frame data output control module; Above-mentioned three modules connect and compose data path successively, and the frame data of frame data output control module output feed back to the interframe data processing module; In the interframe data processing module, carry out the computing between the consecutive frame data, between former frame data and present frame, carry out subtraction, computing gained result multiply by filter factor and adds present frame and obtain filtered current frame data; The frame data cache module is made up of the DDR2 memory, has two buffer areas, and each buffer area is buffer memory one frame video image data respectively; Frame data output after the frame data output control module control Filtering Processing.

The filtering noise-reduction method of video data of the present invention may further comprise the steps:

A, after filtering noise reduction system operation, the input of the 1st frame video image data, filtering this moment noise reduction system is not handled the 1st frame video image data, and directly these frame data is inputed in first buffer area, and the 1st frame video image data are carried out buffer memory;

B, when the 2nd frame video image data are imported, the frame data output control module is outwards exported the 1st frame video image data in first buffer area, uses the 1st frame data of first buffer area and the data of current the 2nd frame to carry out subtraction simultaneously;

C, two frame adjacent video view data multiply by filter factor K, K ∈ [0,1] through the result who draws behind the subtraction;

D, the result who multiply by after the filter factor K add the 2nd current frame video image data, thereby obtain the 2nd frame data after the Filtering Processing, these frame data are stored in second buffer area of DDR2 memory of frame data cache module;

E, when the 3rd frame video image data are imported; The frame data output control module is the 2nd interior frame data of second buffer area of output outwards; Use the 2nd frame data of second buffer area and the 3rd current frame data to carry out subtraction simultaneously; The gained result multiply by filter factor K and draws filtered the 3rd frame data with the 3rd frame data sum again, deposits these data in first buffer area;

F, above-mentioned steps is carried out in circulation when new data are imported once more.

Advantage and good effect that the present invention has are:

In the filtering noise reduction system and filtering noise-reduction method of video data of the present invention; Interframe data processing module, frame data cache module and frame data output control module are linked to be data path, between former frame data and present frame, carry out subtraction, computing gained result multiply by filter factor and adds present frame and obtain filtered current frame data; Thereby can adapt to the denoising of the video image of low speed mobile object; Do temporal weighted average to picture signal with the frame period, then, promptly can keep the edge in the image well almost not influence of picture signal; Can reduce noise greatly, thereby reach the purpose of improving signal noise ratio (snr) of image.

Description of drawings

Fig. 1 is the structural representation of the filtering noise reduction system of video data of the present invention.

Embodiment

Followingly the present invention is carried out detailed explanation with reference to accompanying drawing and embodiment.

Fig. 1 is the structural representation of the filtering noise reduction system of video data of the present invention.

As shown in Figure 1, the filtering noise reduction system of video data of the present invention comprises interframe data processing module, frame data cache module and frame data output control module, and the wr_ctrl module between each module is an input control module, and rd_ctrl is an output control module; Above-mentioned three modules connect and compose data path successively, and the frame data of frame data output control module output feed back to the interframe data processing module; In the interframe data processing module, carry out the computing between the consecutive frame data, between the Frame pxd_i of former frame data and current input, carry out subtraction, computing gained result multiply by filter factor and adds present frame and obtain filtered current frame data; The frame data cache module is made up of the DDR2 memory, has two buffer area FIFO1 and FIFO2, and each buffer area is buffer memory one frame video image data respectively; Frame data output pxd_o after the frame data output control module control Filtering Processing.

In digital camera system, repeat by the frame period by the dynamic digital image of digital camera collection, except that the image that switches at a high speed, the coefficient correlation between two frames can reach 0.8, and it is identical that the image information of promptly adjacent two frames has 80 %.Be mingled in noise in the image except that dark current noise, most noise is at random.Do temporal weighted average to picture signal with the frame period, then, promptly can keep the edge in the image well, can reduce noise greatly, thereby reach the purpose of improving signal noise ratio (snr) of image almost not influence of picture signal.

This mathematic(al) representation is:

I_out?=?I_i?+?K?_*?(I₀-I_i)。

I wherein₀: the former frame image; I_i: current frame image; I_Out: filtered present frame; K: filter factor, K ∈ [0,1].

The filtering noise-reduction method of video data of the present invention may further comprise the steps:

A, after filtering noise reduction system operation, the input of the 1st frame video image data, filtering this moment noise reduction system is not handled the 1st frame video image data, and directly these frame data is inputed in first buffer area, and the 1st frame video image data are carried out buffer memory;

B, when the 2nd frame video image data are imported, the frame data output control module is outwards exported the 1st frame video image data in first buffer area, uses the 1st frame data of first buffer area and the data of current the 2nd frame to carry out subtraction simultaneously;

C, two frame adjacent video view data multiply by filter factor K, K ∈ [0,1] through the result who draws behind the subtraction;

D, the result who multiply by after the filter factor K add the 2nd current frame video image data, thereby obtain the 2nd frame data after the Filtering Processing, these frame data are stored in second buffer area of DDR2 memory of frame data cache module;

E, when the 3rd frame video image data are imported; The frame data output control module is the 2nd interior frame data of second buffer area of output outwards; Use the 2nd frame data of second buffer area and the 3rd current frame data to carry out subtraction simultaneously; The gained result multiply by filter factor K and draws filtered the 3rd frame data with the 3rd frame data sum again, deposits these data in first buffer area;

F, above-mentioned steps is carried out in circulation when new data are imported once more.

The above only is preferred embodiment of the present invention, is not the present invention is done any pro forma restriction; Though the present invention is with preferred embodiment openly as above, yet, be not in order to limit the present invention; Anyly be familiar with the professional and technical personnel, in not breaking away from technical scheme scope of the present invention, can utilize the technology contents of announcement to make a little change or modification certainly; Become the equivalent embodiment of equivalent variations; In every case be the content that does not break away from technical scheme of the present invention, to any simple modification, equivalent variations and modification that above embodiment did, all belong in the scope of technical scheme of the present invention according to technical spirit of the present invention.

Claims (2)

1. the filtering noise reduction system of a video data; It is characterized in that: comprise interframe data processing module, frame data cache module and frame data output control module; Above-mentioned three modules connect and compose data path successively, and the frame data of frame data output control module output feed back to the interframe data processing module; In the interframe data processing module, carry out the computing between the consecutive frame data, between former frame data and present frame, carry out subtraction, computing gained result multiply by filter factor and adds present frame and obtain filtered current frame data; The frame data cache module is made up of the DDR2 memory, has two buffer areas, and each buffer area is buffer memory one frame video image data respectively; Frame data output after the frame data output control module control Filtering Processing.

2. filtering noise-reduction method based on the filtering noise reduction system of the described video data of claim 1 may further comprise the steps:

A, after filtering noise reduction system operation, the input of the 1st frame video image data, filtering this moment noise reduction system is not handled the 1st frame video image data, and directly these frame data is inputed in first buffer area, and the 1st frame video image data are carried out buffer memory;

B, when the 2nd frame video image data are imported, the frame data output control module is outwards exported the 1st frame video image data in first buffer area, uses the 1st frame data of first buffer area and the data of current the 2nd frame to carry out subtraction simultaneously;

C, two frame adjacent video view data multiply by filter factor K, K ∈ [0,1] through the result who draws behind the subtraction;

D, the result who multiply by after the filter factor K add the 2nd current frame video image data, thereby obtain the 2nd frame data after the Filtering Processing, these frame data are stored in second buffer area of DDR2 memory of frame data cache module;

E, when the 3rd frame video image data are imported; The frame data output control module is the 2nd interior frame data of second buffer area of output outwards; Use the 2nd frame data of second buffer area and the 3rd current frame data to carry out subtraction simultaneously; The gained result multiply by filter factor K and draws filtered the 3rd frame data with the 3rd frame data sum again, deposits these data in first buffer area;

F, above-mentioned steps is carried out in circulation when new data are imported once more.

Images