CN101106699A

CN101106699A - Method for deciding kept video in monitoring device

Info

Publication number: CN101106699A
Application number: CNA2006101014844A
Authority: CN
Inventors: 徐纬
Original assignee: Primax Electronics Ltd
Current assignee: Primax Electronics Ltd
Priority date: 2006-07-10
Filing date: 2006-07-10
Publication date: 2008-01-16

Abstract

The invention provides a method to determine the reserved images in monitoring devices. The invention obtains a plurality of image files by using the photography device in a monitoring device to continuously photo focused images, compares the file size of each image file with a critical value of one file size, and selects image files to be reserved from a plurality of the image files. The invention doesn't need to analyze the definition values of images, and determines whether to reserve images only by comparing file sizes; without need to calculate definition values of images, to determine whether to reserve images takes less time and can also save the resources of the monitoring devices.

Description

The method of decision kept video in supervising device

Technical field

The present invention relates to a kind of supervising device, especially relate to a kind of method that in supervising device, determines kept video.

Background technology

Supervising device is generalizing in the society now gradually, from previous bank, national treasury, post office, ATM (automatic teller machine), hospital and prison etc., and some deposit money, look after sufferer and guard prisoners etc. and need the place that monitors, and convenience store till now, company, building, sub-district, elevator and parking lot etc. almost are that any place all has supervising device.

Because supervising device is day by day popularized, so the development of supervising device is also more and more progressive.Known supervising device roughly is divided into two kinds, a kind ofly is to use the mode of direct video recording and does not carry out the action of compressing image file.Another kind of then be supervisory control system with blurring checkout gear.

First kind of known supervising device of directly recording a video adopts the mode that continues video recording, that is to say, all images that photograph are all kept file, when needs take out specific picture, from huge document retaining or video tape, search one by one again, very trouble and consuming time.

Second kind of supervising device then is provided with image movement of objects checkout gear, and this movement of objects checkout gear can judge whether that object is moving, and then judges whether the image of being recorded should keep.Its reason is, when there are objects moving in the scene that is taken, represents that monitored place has the situation incident to take place, and captured in such cases image is only the object of the required reservation of supervising device.The function mode of this kind supervising device can be saved a large amount of internal memories.

The enforcement principle of supervising device with blurring device is as follows: an object (this object can be the object that chair, chest etc. can be focused) is placed in the position that the camera in the supervising device is selected to be positioned between camera and the object (object is ATM (automatic teller machine), safety box, desk etc.), camera is focused to this object, and wherein the position at this object place is a focusing.After focusing is finished this object removed away from coverage and according to this focal length and take continuously,, around this focusing, have a depth of field (Depth of Field) according to optical principle.The characteristic of the depth of field is that so long as the focusing of being carried out is taken, the image that is obtained is sharp image in field depth.Because camera is not to focus on object and nobody or object are positioned at field depth, therefore, this moment, the definition values (Focus Value) of image was very little, that is to say that image is fuzzy.When people or object entered in the field depth of camera, it is clear that the captured image of camera this moment becomes, and promptly the definition values of image becomes big, and supervising device to be variation with definition values judge whether to keep image.

Yet, second kind of known process need that keeps captured image that judges whether in supervising device decompresses to each image file earlier, analyze this image again to calculate the definition values of image, whether clear and whether should keep this image according to this image of the definition values section of declaring that is obtained again.Though comparing with first kind of supervising device, second kind of known supervising device have the advantage of saving internal memory, yet the defective of second kind of known supervising device then is, when camera is taken, supervising device carries out the image analysing computer action simultaneously, may cause supervising device system resource overload under these circumstances, cause cost to increase.

Summary of the invention

The object of the present invention is to provide a kind of in supervising device the method for decision kept video, the size by using image file is as the basis of definition judgment, can effectively shorten the time that judges whether to keep image under the situation of consumption of natural resource system not.

For achieving the above object, the invention provides a kind of method that determines kept video in supervising device, this supervising device comprises camera, and this camera obtains a plurality of image files in order to the image of continuous shooting focusing position, and this method comprises:

The file size of each image file is compared with the file size critical value, with the image file of from a plurality of image files, selecting to answer monitored device to keep.

In a preferred embodiment, wherein:

When the file size of this image file during, keep a plurality of image files before this image file and this image file more than or equal to this document size critical value; Or

When the file size of this image file during more than or equal to this critical value, keeps this image file less than the size of a plurality of image files before this critical value and this image file.

Preferably, wherein this image file is suppressed by the Frequency Domain Coding compression algorithm and is formed.

Preferably, wherein this image file is jpeg file (Joint Photographic Experts Groupfile).

Compare with known method, the present invention does not need to analyze the definition values of image, as long as the comparison document size can determine whether keep image,, therefore judge whether to keep the resource that the required time of image is short and can save supervising device owing to needn't calculate the definition values of image.

Description of drawings

Fig. 1 is the enforcement illustration of a plurality of images of output of the present invention.

Fig. 2 is the curve chart of correlation between the file size of captured image of signal the present invention and described image.

Wherein, description of reference numerals is as follows:

TH image file size critical value

A, B, C, D image file size are less than the image of image file size critical value TH

E, F file size equal the image of image file size critical value TH value

Embodiment

The present invention is identical haply with the process of known method, and main difference is on the Study document that known method determines whether keeping image for calculating definition values, and the present invention is for directly to judge with the size of image file.

Hereinafter basic principle of the present invention will be described earlier.

General image file has multiple file format, and for example: BMP, JPEG, TIF, GIF and PNG or the like, the Frequency Domain Coding method is used in present multimedia compression mostly, for example illustrates with jpeg file.When image capture unit obtains image by taking, the image that is untreated of this moment is called initial data (Raw data), transfer the image of handling to through some image processing (Image pipeline), be called the YCbCr image, it is the jpeg file file format that method by a kind of Frequency Domain Coding makes the YCbCr video conversion, at first carry out discrete cosine transform (Discrete Cosine Transform, DCT), make image be converted to frequency domain by time domain, quantize (Quantization) again, huffman coding method (Huffman coding) and packing document head (Header) supervisor can be jpeg file with the initial data video conversion.

In this kind Frequency Domain Coding method, carry out after the huffman coding method, the Frequency Domain Coding length of image is determined, crucial part of the present invention is Frequency Domain Coding length, the length of coding is represented the complexity of image, and image is complicated more more clearly, that is to say that Frequency Domain Coding length is long more, and image is just clear more.And the action that after the huffman coding method is finished, must pack the document head, the file size of document head is fixed, and is added in just to make it become complete jpeg file on the coding originally and can not have influence on relative file size relation.Therefore can utilize the size of image file to decide the whether clear of image.

In the method for the invention, at first set the file size critical value TH of an image, this file size critical value TH is the file size of image in clear boundary.The file size of supposing captured image is Fs.When Fs represents that image is clear during more than or equal to TH, Fs represents that then image is fuzzy during less than TH.

The present invention judges that the condition that keeps image is: as the big or small Fs of image file during more than or equal to file size critical value TH, keep first three image file of opening of this image file and this image file; Or when the size of the 3rd image file is more than or equal to critical value TH before the big or small Fs of image file is less than critical value TH and image file, keep this image file.

See also Fig. 1, it takes the focusing a plurality of image outputs that the position obtained for supervising device of the present invention is continuous embodiment comprises 34 figure such as Fig. 1 .1 to Fig. 1 .34 altogether.The picture of Fig. 1 .1 to Fig. 1 .13 presents identical content, represents that any change does not take place monitored scene, and just without any object is invaded monitored zone.Then all show the not image in the monitoring area of script in the image of Fig. 1 .14 to Fig. 1 .27, expression has people/object invasion monitoring area.At last, Fig. 1 .28 to Fig. 1 .34 shows again and original monitored regional identical presentation content, and the expression invador has left monitored zone.

See also Fig. 2, it is the curve chart of correlation between the file size of the captured a plurality of images of signal the present invention and described a plurality of images.Wherein the longitudinal axis is a file size, and transverse axis is an image sequence.Please consult Fig. 1 and Fig. 2 simultaneously, the TH among Fig. 2 represents the critical value of predefined image file size.By beginning to take file size to E point (not comprising the E point) all less than TH, represent that promptly the image of this section all blurs, contrast Fig. 1 is by Fig. 1 .1 to Fig. 1 .15 that begins to take to E point (the not comprising the E point) presentation graphs 1.Get back to Fig. 2, wherein the expression of B point has the people to enter coverage, because people or object in moving are difficult to focus for camera, therefore people or the object taken in moving can obtain fuzzy image, B point among Fig. 2 is the minimum point of preceding half section curve, contrast Fig. 1, the image that B is ordered is Fig. 1 .14.The Fs of Fig. 2 when the E point equals TH, the people that expression enters coverage has rested in the coverage and image begin to become clear, so the image when camera is retained in the E point.The Fs of ordering to F after the E point is all more than or equal to TH, so the image of EF section also all is retained.Contrast Fig. 1, the image that E is ordered is Fig. 1 .16, and the image that F is ordered is Fig. 1 .23.

Supposed only to keep the image of the EF section among Fig. 2, Fig. 1 .16 to Fig. 1 .23 just, before then can seeing that the people who enters coverage suddenly appears at camera lens playing video the time, and the invador just prepares to exit but and does not also leave when finishing, and can cause the invador that reflects that the image that is retained can not be complete to enter and leave the monitoring area from that direction like this.For keeping the complete image record that the invador enters the monitoring area really, therefore when keeping E point image, three image also must be preserved before the E point, just begins to keep image from the A point, and also need to keep at back three images that F is ordered, just remain into the D point.Therefore the image that in fact needs to keep is the AD section, has comprised the people who enters coverage and has begun just the complete image of image that enters coverage (B point) and leave (C point) fully.In Fig. 1, the image that A is ordered promptly is Fig. 1 .13, and the image that C is ordered promptly is Fig. 1 .25, and the image that D is ordered promptly is Fig. 1 .27, so the image that all is retained is Fig. 1 .13 to Fig. 1 .27.

The above is preferred embodiment of the present invention only, is not in order to limiting scope of patent protection of the present invention, and therefore every other do not break away from being equal to of being finished under the spirit disclosed in this invention and change or modify, and all should be contained in the application's the scope of patent protection.

Claims

1. the method for a decision kept video in supervising device, this supervising device comprises a camera, and this camera obtains a plurality of image files in order to the image of continuous shooting one focusing position, and this method comprises:

The file size of each this image file is compared with a file size critical value, with the image file of from described a plurality of image files, selecting to be kept by this supervising device.

2. the method that in supervising device, determines kept video as claimed in claim 1, wherein:

When the file size of described image file during, keep a plurality of image files before this image file and this image file more than or equal to this document size critical value; Or

When the file size of described image file during more than or equal to this critical value, keeps this image file less than the size of a plurality of image files before this critical value and this image file.

3. the method that determines kept video in supervising device as claimed in claim 2, wherein this image file is suppressed by the Frequency Domain Coding compression algorithm and is formed.

4. the method that determines kept video in supervising device as claimed in claim 3, wherein this image file is a jpeg file.