CN1549204A - Method for detecting and reconstructing dynamic image pixel - Google Patents

Abstract

The invention discloses a method for detecting and reconstructing dynamic image pixels, and provides a method for reconstructing and calculating missing pixels in an image converted from interlaced scanning to progressive scanning by using a static counting graph. Firstly, whether the missing pixel is located in a static area or a non-static area is judged, and the purpose of optimally reconstructing the missing pixel is achieved by a front-back field interpolation method and a peripheral field interpolation method respectively. The reconstruction steps comprise: inputting the time field and a plurality of reference fields; scanning out the missing pixels in the time field in the image; calculating a difference value of pixels surrounding the missing pixel; comparing the difference value with a critical value; correcting the value of a static counter according to the comparison result of the difference value and the critical value; reconstructing the missing pixel by using a front-to-back field interpolation method or a peripheral field interpolation method according to the value of the static counter; outputting the reconstructed deinterlaced pixels; and calculating whether the scanning and reconstructing steps are completed.

Description

The detection of dynamic image pixel and method for reconstructing

Technical field

The present invention relates to a kind of detection and method for reconstructing of dynamic image pixel, propose a kind of with static counting diagram (Static counter map) as the reconstruction determination methods of omitting pixel (missing pixels) in interlacing (interlaced-scan) image, judge that this pixel is that to be positioned at static region (staticregion) also be non-static region (non-static region), reach with interpolation method again and rebuild the purpose of omitting pixel.

Background technology

It also is non-static region that a kind of dynamic adjustment algorithm of the prior art (motion-adaptive algorithm) is positioned at static region with the difference between the adjacent perimeter pixel as omission pixel in the judgement interlacing image.Be positioned at static region if judge this omission pixel, just rebuild this omission pixel with the front and back interpolation field method (inter-field interpolation) of utilizing contiguous (neighboring fields) pixel information; Be positioned at non-static region if judge to omit pixel, then the peripheral interpolation field method (intra-field interpolation) of the former sweep trace of vicinity (the neighboring original scan lines) information by scanning field (scanfield) is rebuild this omission pixel.

Front and back interpolation field method synoptic diagram with reference to Figure 1A prior art.Wherein omitting pixel 10 betides in the pixel that P-SCAN (progressive-scan) abandoned in the interlacing conversion, X as shown in the figure, and be expressed as zero, the preceding time pixel 11 of this omission pixel 10 is respectively time field F (n-1) and the middle pixel that exists originally of F (n+1) with back time pixel 13, the coordinate that its each time is gone up this pixel is (x, y), preceding time pixel 11, omit pixel 10 and back time pixel 13 respectively three different demonstration times, if will rebuild this omission pixel 10, the back court interpolation method is preceding time pixel 11 averaged with the image information of back time pixel 13 and obtains omitting the reconstructed value of pixel 10 before this, that is:

$X=\frac{[F(x,y,n-1)+F(x,y,n+1)]}{2}---\left(1\right)$

Wherein, (x, y n-1) are the pixel function of preceding pixel 11 to F, and (x, y n+1) are the pixel function of back time pixel 13 to F.

Peripheral interpolation field method synoptic diagram with reference to prior art among Figure 1B.Wherein omit pixel 10 and be positioned at the X shown in the figure, and be depicted as zero, primary importance pixel 14 and the pixel of second place pixel 16 for existing originally, primary importance pixel 14, omission pixel 10 are respectively (x with the coordinate of second place pixel 16, y-1), (x, y) with (x, y+1), be three demonstrations of pixel up and down of same time field F (n), omit pixel 10 if will rebuild, with peripheral interpolation method primary importance pixel 14 being averaged with the image information of second place pixel 16 obtains omitting the reconstructed value of pixel 10, that is:

$X=\frac{[F(x,y-1,n)+F(x,y+1,n)]}{2}---\left(2\right)$

Wherein, (x, y-1 n) are the pixel function of primary importance pixel 14 to F, and (x, y+1 n) are the pixel function of second place pixel 16 to F.

It also is non-static region that desire decision omission pixel 10 is in static state, can utilize around this omission pixel 10 information gap of adjacent pixel to judge, if difference value is less than a critical value (threshold), pixel judges then that for static this omission pixel 10 is for static around representing this omission pixel 10; If difference value is more than or equal to this critical value, then pixel is non-static state all around, judges that promptly this omission pixel 10 is positioned at the dynamic area.

With reference to Fig. 2 calculated difference value of the prior art method synoptic diagram.This figure is depicted as the computing method of above-mentioned difference value, utilize shown in the figure the preceding time field 21 and peripheral absolute value summation (the sum ofabsolute difference of the pixel calculated difference of existence originally that shown in the time field 23 of back is zero of the omission pixel 20 that is X, SAD), be about to the front and back interpolation method of above-mentioned Figure 1A surrounding time field and the peripheral interpolation method join operation of Figure 1B upper-lower position, its calculation formula is as follows:

$\mathrm{Diff}(x,y,n)=\underset{(i,j)\∈\mathrm{\Γ}}{\mathrm{\Σ}}|f(i,j,n-1)-f(i,j,n+1)|---\left(3\right)$

Γ＝{(x，y-2)，(x，y)，(x，y+2)，(x-1，y)，(+1，y)} (4)

Wherein (x, y n) are the difference value functional expression to Diff, ∑ | and f (i, j, n-1)-f (i, j, n+1) | be the absolute value summation of the peripheral pixel calculated difference of omitting pixel 20 surrounding time fields, function f (i, j n-1) is illustrated in preceding time field 21 shown in Figure 2, and function is each the pixel function in the F (n-1), the information that comprises each pixel, f (i, j, n+1) then be illustrated in back time field 23, function is each the pixel function in the F (n+1), and (i j) then represents the position of surrounding time field pixel, the position includes (x, y-2), (x, y), (x, y+2), (x-1, y) with (x+1, y).

Fig. 3 is the dynamic pixel adjustment calculation process flow diagram of prior art.As shown in the figure, F (n) is a time field function, when dynamically pixel adjustment calculation begins, this time field function F (n) is calculated 301, one release of an interleave processor (deinterlacing processor) field F input time (n), preceding time field F (n-1) and the back time field 303 that function is F (n+1) by release of an interleave (deinterlaced).Afterwards, scan the omission pixel among the time field F (n) in the image, and carry out interpolation 305 in proper order at a grating, upper left by image just to the bottom right P-SCAN, afterwards with the described calculation formula of Fig. 2 ring around the difference value 307 of omitting pixel around the pixel, and judge relatively that by this and critical value omitting pixel is to be positioned at static or non-static region 309, if difference value Diff is less than this critical value, just rebuild 311 with the described front and back of Figure 1A interpolation field method, if difference value Diff is more than or equal to critical value, then rebuild 313 with the described peripheral interpolation field method of Figure 1B, export the release of an interleave pixel of rebuilding 315 again, judge whether to finish the release of an interleave calculation 317 of this field function F (n) afterwards, if not, then continue scanning and omit pixel, and carry out interpolation 305 at a grating in proper order, the omission pixel is rebuild fully in time field F (n); If then expression finishes the release of an interleave calculation 319 of this field function F (n).

The above is a dynamic pixel adjustment calculation flow process in the prior art, the method is easily implemented simply again, but because only judge that with the difference value of a field omitting pixel is positioned at static or non-static region, though can be used for the dynamically judgement of pixel of zone, but can make false judgment to quick acting or unclean action into static region, this is because the difference between contiguous can be because of quick acting or ambiguous being judged as less than critical value, and interpolation field method and the omission pixel that makes the mistake is rebuild before and after using.Do not have too much influence if having only one or two omission pixel to rebuild, just if having a lot of pixels that need rebuild can cause the image distortion of certain degree because of false judgment.

Rebuild for solving the above-mentioned mistake that causes image to omit pixel because of false judgment, detection and method for reconstructing that the present invention proposes a kind of dynamic image pixel reach correct static state and the judgement of non-static region.

Summary of the invention

The present invention is about a kind of detection and method for reconstructing of dynamic image pixel, proposition is with the reconstruction computing method of a kind of static counting diagram as omission pixel in the image, with the quiescent state counter contiguous peripheral information of omitting pixel that adds up, correct judge that this pixel is that to be positioned at static region also be non-static region, and reach the purpose of optimal reconstruction omission pixel respectively with peripheral interpolation field method and front and back interpolation field method.

For reaching above-mentioned purpose, the detection of dynamic image pixel of the present invention and method for reconstructing utilize one of release of an interleave calculating to comprise the time field of omitting pixel, in a release of an interleave processor input time field, import a previous time field and a back time field in addition, and should omit picture element scan and went out, ring is omitted a pixel difference value of pixel all around around this again, the size of this difference value and a critical value relatively, and back court interpolation method or rebuild the omission pixel before deciding with the value of quiescent state counter with peripheral interpolation field method, export the release of an interleave pixel of rebuilding afterwards again, judge whether to finish the release of an interleave calculation of this time field at last, rebuild up to whole omission pixels.

Description of drawings

Figure 1A is the front and back interpolation field method synoptic diagram of prior art;

Figure 1B is the peripheral interpolation field method synoptic diagram of prior art;

Fig. 2 is the calculated difference value method synoptic diagram of prior art;

Fig. 3 is the dynamic pixel adjustment calculation process flow diagram of prior art;

Fig. 4 uses the dynamic pixel adjustment calculation process flow diagram of static counting diagram for the embodiment of the invention;

Fig. 5 A is field, the static counting diagram upper strata synoptic diagram of the embodiment of the invention;

Fig. 5 B is the static counting diagram lower floor field synoptic diagram of the embodiment of the invention;

Fig. 5 C is the static counting diagram synoptic diagram of the embodiment of the invention;

Fig. 6 is that the difference value of the embodiment of the invention is calculated synoptic diagram;

Fig. 7 is the border of the embodiment of the invention peripheral shadow lattice interpolation method synoptic diagram that leads;

Fig. 8 is the dynamic image pixel adjustment calculation process flow diagram of the embodiment of the invention.

Description of reference numerals

10 omit pixel 11 preceding time pixels

13 back time pixel 14 primary importance pixels

16 second place pixels 20 are omitted pixel

Time field 23 back time field before 21

51 former sweep trace 53 former scanning elements

55 sweep traces, 57 scanning elements

501 first omit pixel 503 second omits pixel

61 very first time fields, 63 second time field

65 the 3rd time field 67 the 4th time field

70 omit pixel 71 first adjacent pixels

72 second adjacent pixels 73 the 3rd adjacent pixel

Nearly pixel 75 the 5th adjacent pixel of 74 neighbours

76 the 6th adjacent pixel D1, first difference value

D2 second first critical value of difference value Th1

Th2 second critical value

Embodiment

Use the dynamic pixel adjustment of static counting diagram to calculate flow process with reference to Fig. 4 embodiment of the invention, propose static counting diagram as judging basis static and non-static region.As shown in the figure, F (n) is a time field function, when this dynamic pixel adjustment calculation flow process begins, this time field function F (n) is calculated (step 401) by release of an interleave, the release of an interleave processor is imported a time field F (n), with a plurality of reference field, is the back time field (step 403) of F (n+1) as preceding time field F (n-1) with a function, afterwards, go out the omission pixel among the time field F (n) in the image with the grating P-SCAN, and carry out interpolation (step 405), this omissions pixel betides P-SCAN in the pixel that is abandoned during interlacing is changed.With the described difference value calculation of Fig. 2 formula, be formula (3), ring is around omitting the pixel difference value (step 407) of pixel all around, and judge relatively that by this and critical value omitting pixel is to be positioned at static or non-static region (step 409), and according to the value of the comparative result correction quiescent state counter of this difference value and critical value, if difference value is less than this critical value, judge that then this omission pixel is positioned at static region, just added value 1 is to the value (step 411) of quiescent state counter, and accumulate this static counting value (step 413) with aforementioned judgement, when this static counting value less than fixed counting critical value, still rebuild non-static peripheral interpolation field method of omitting pixel and rebuild this omission pixel (step 419) with prior art; When this static counting value was accumulated to more than or equal to the counting critical value, just the back court interpolation method was rebuild and is omitted pixel (step 415) in the past.When ring around the difference value of omitting pixel around the pixel more than or equal to critical value (step 409), judge that then this omission pixel is positioned at the dynamic area, remove the value (step 417) of quiescent state counter, making it is 0, and continues to rebuild omission pixel (step 419) with the described peripheral interpolation field method of Figure 1B.Export the release of an interleave pixel of rebuilding (step 421) again, afterwards, calculate whether to scan and finish with reconstruction procedures, to judge whether to finish the release of an interleave calculation (step 423) of this field function F (n), if not, then continue scanning and omit pixel and carry out interpolation (step 405) at grating in proper order, the omission pixel is rebuild fully in time field F (n); If then expression finishes the release of an interleave calculation (step 425) of this field function F (n).

Static counting diagram running synoptic diagram shown in Fig. 5 A to Fig. 5 C being the embodiment of the invention wherein is divided into image frame respectively field, upper strata (top-field) and lower floor field (bottom-field).Shown in the static counting diagram lower floor field synoptic diagram of the static counting diagram upper strata field synoptic diagram of Fig. 5 A embodiment of the invention and Fig. 5 B embodiment of the invention, it is former sweep trace 51 (original scan lines) that black part is divided the strip rectangle that forms, and black box is former scanning element 53 (original pixels), and white portion then is respectively sweep trace 55 (missing scan lines) and scanning element 57 (missing pixels).Symbol X is depicted as the rebuilt first omission pixel 501 and second and omits pixel 503, Fig. 5 C is the static counting diagram synoptic diagram of the embodiment of the invention, wherein symbol X is depicted as the omission pixel that adds up, and static counting diagram first of the field, Fig. 5 A upper strata of adding up is omitted pixel 501 and omitted pixel 503 with second of lower floor field shown in Fig. 5 B among the figure.

Judge that for improving the omission pixel is positioned at the accuracy of static or non-static region, the present invention calculates the difference value of field, upper strata and lower floor's field image simultaneously.Fig. 6 is that the difference value of the embodiment of the invention is calculated synoptic diagram.The present invention uses four fields to come the calculated difference value in image frame, is respectively very first time field 61, and function is F (n-2), second time field 63, function are F (n-1), and this omits the 3rd time field 65 at pixel place, function is F (n), the 4th time field 67, and function is F (n+1).Wherein need by the field of release of an interleave for the 3rd time field 65 of omitting pixel is arranged, all the other are the required reference field of calculated difference value, with the 3rd time field 65 is benchmark, just omit the reference field at pixel place, as shown in Figure 6, very first time field 61 is again the time field of preceding state, with the 3rd time field 65 time field of same scan state is arranged, second time field 63 be before the time field of state, the 4th time field 67 be after the time field of state, and this second time field 63 and the 4th time field 67 are the time field of the scanning mode different with the 3rd time field 65.And calculate the first difference value D1 by each pixel location of very first time field 61 and the 3rd time field 65; Calculate the second difference value D2 by second time field 63 and the 4th time field 67 each pixel location, its calculation formula is as follows respectively:

$D1(x,y,n)=\underset{(i,j)\∈\mathrm{\Γ}1}{\mathrm{\Σ}}|f(i,j,n)-f(i,j,n-2)|---\left(5\right)$

$D2(x,y,n)=\underset{(i,j)\∈\mathrm{\Γ}2}{\mathrm{\Σ}}|f(i,j,n-1)-f(i,j,n+1)|---\left(6\right)$

Wherein Γ 1 is the position of pixel with Γ 2, and is as follows respectively:

Γ1＝[(x-1，y-1)，(x，y-1)，(x+1，y-1)，(x-1，y+1)，(x，y+1)，(x+1，y+1)] (7)

Γ2＝[(x，y-2)，(x，y)，(x，y+2)，(x-1，y)，(x+1，y)] (8)

The very first time field 61 in the pixel location representative graph 6 of above-mentioned formula (7) and the pixel location of the 3rd time field 65, and second time field 63 in the pixel location representative graph 6 of formula (8) and the pixel location of the 4th time field 67 calculate the otherness of this omission pixel periphery field again by formula (5) and formula (6).

Fig. 7 is the border of the embodiment of the invention peripheral shadow lattice interpolation method synoptic diagram that leads.This border peripheral interpolation method that leads is selected the function difference reckling of the peripheral diagonal position pixel of this omission pixel, and then the mean value of this diagonal position pixel is the reconstructed value of this omission pixel.As shown in the figure, symbol X represents that is omitted a pixel 70, its contiguous pixel has first adjacent pixel 71, second adjacent pixel 72, the 3rd adjacent pixel 73, the nearly pixel 74 of neighbours, the 5th adjacent pixel 75 and the 6th adjacent pixel 76, utilizes between the corresponding adjacent pixel otherness to judge that this omissions pixel 70 is static state or non-static state.

Wherein, the first difference U1 is expressed as the difference of diagonal position first adjacent pixel 71 and the 6th adjacent pixel 76; The second difference U2 is expressed as the difference of diagonal position second adjacent pixel 72 and the 5th adjacent pixel 75; The 3rd difference U3 represents the difference of the 3rd adjacent pixel 73 and the nearly pixel 74 of neighbours, and finds out the criterion that wherein minimum difference is used as static pixel and non-static pixel.If minimum value is U1, as shown in the figure, then omit the mean value that pixel 70 is first adjacent pixel 71 and the 6th adjacent pixel 76; If minimum value is U2, then omit the mean value that pixel 70 is second adjacent pixel 72 and the 5th adjacent pixel 75; If minimum value is U3, then omit the mean value that pixel 70 is the 3rd adjacent pixel 73 and the nearly pixel 74 of neighbours.Further obtain omitting the reconstructed value of pixel 70 afterwards again with the intermediate value filtration method, this intermediate value filtration method is for limiting this omission pixel reconstructed value error, and find out the method for the intermediate value of this omission pixel and relative pixel up and down as this omission pixel reconstructed value, present embodiment is the intermediate value of finding out with second adjacent pixel 72, the 5th adjacent pixel 75, with this intermediate value as omitting the new reconstructed value of pixel 70.The above is the lead implementation content of peripheral shadow lattice interpolation method (edge-oriented intra-frame interpolation) and intermediate value filtration method (median filtering) of border.

Dynamic image pixel adjustment calculation process flow diagram with reference to Fig. 8 embodiment of the invention, the first difference value D1 and the second difference value D2 that calculate with the otherness between each shown in Figure 6, and set two corresponding critical values and judge that omitting pixel is static also right and wrong static state.As shown in the figure, F (n) is a time field function, when this dynamic pixel adjustment calculation flow process begins, this time field function F (n) is calculated (step 801) by release of an interleave, the release of an interleave processor is imported each time field function (step 803), be respectively very first time field 61, function is F (n-2), second time field 63, function is F (n-1), this omits the 3rd time field 65 at pixel place, function is that F (n) and function is the 4th time field 67 of F (n-1), afterwards, scan in the image omission pixel among the time field F (n) and carry out interpolation (step 805) in proper order at grating, with the described difference value calculation of Fig. 2 formula, and by each interfield disparity calculated difference value (step 807) shown in Figure 6, be respectively that very first time field 61 and the 3rd time field 65 as formula (5) calculates the first difference value D1, calculate the second difference value D2 with second time field 63 and the 4th time field 67 as formula (6), set the corresponding first critical value Th1 and the second critical value Th2 again, and judge relatively that by these two critical values and two difference value omitting pixel is to be positioned at static or non-static region (step 809).

Afterwards, if the first difference value D1 less than this first critical value Th1 and the second difference value D2 less than the second critical value Th2, judge that then this omission pixel is positioned at static region, added value 1 is in the value (step 811) of quiescent state counter, and with this judge the accumulation this static counting value (step 813), when this static counting value less than fixed counting critical value, rebuild this omission pixel (step 819) with the border shown in Figure 7 peripheral shadow lattice interpolation method that leads, when this static counting value is accumulated to more than or equal to the counting critical value, then rebuilds and omit pixel (step 815) with the front and back interpolation field method of prior art.

Judging when omitting the difference (step 809) of pixel difference value and critical value around the pixel, as the first difference value D1 more than or equal to first critical value Th1 or the second difference value D2 more than or equal to the second critical value Th2, judge that then omitting pixel is positioned at the dynamic area, just remove the value (step 817) of quiescent state counter, make it is 0, and continue to rebuild this omissions pixel (step 819) with the border shown in Figure 7 peripheral shadow lattice interpolation method that leads, follow the reconstruction (step 821) that obtains omitting pixel with the intermediate value filtration method.Export the release of an interleave pixel of rebuilding (step 823) again, judge whether to finish the release of an interleave calculation (step 825) of this field function F (n) afterwards, if not, step continues scanning and omits pixel and carry out interpolation (step 805) at grating in proper order before then getting back to, and the omission pixel is rebuild fully in time field F (n); If then expression finishes the release of an interleave calculation (step 827) of this field function F (n).

More than be the detailed description of the detection of dynamic image pixel of the present invention and method for reconstructing embodiment, the present invention improves omitting the judgement of pixel static state and non-static state by static counting diagram, omit pixel in order to correct reconstruction, use lead peripheral shadow lattice interpolation method and intermediate value filtration method of border to carry out stricter reconstruction in addition.

Claims (19)

1. A method for detection and reconstruction of dynamic image pixels, used for reconstructing missing pixels, the missing pixels are located in a time field, and the reconstruction steps include: Input the time field and multiple reference fields; Scanning out the missing pixel in the time field of the image; Compute a disparity value for surrounding pixels around the missing pixel; Comparing the difference value with a critical value; Modifying the value of a static counter, according to the comparison result of the difference value and the critical value; Reconstructing the missing pixel, reconstructing the missing pixel according to the value of the static counter by front and back field interpolation or peripheral field interpolation; output reconstructed de-interlaced pixels; and Calculates whether the scan and rebuild steps are complete. 2. The method for detection and reconstruction of dynamic image pixels as claimed in claim 1, wherein the missing pixels are pixels discarded during sequential scan to interlaced scan conversion. 3. The method for detection and reconstruction of dynamic image pixels as claimed in claim 1, wherein the reference field is a front time field and a back time field of the time field. 4. The method for detecting and reconstructing dynamic image pixels as claimed in claim 1, wherein the difference value is the absolute value of the difference between the previous time field of the missing pixel and a plurality of surrounding pixels of the back time field sum. 5. The detection and reconstruction method of dynamic image pixels as claimed in claim 1, wherein if the difference value is less than the critical value, the value of the static counter is increased by 1, and if the difference value is greater than or equal to the critical value, then Clear the value of this static counter to 0. 6. The method for detection and reconstruction of dynamic image pixels as claimed in claim 1, wherein if the value of the static counter is greater than or equal to a count threshold value, then the missing pixel is reconstructed by front and back field interpolation, if the difference If the value is less than the count threshold, the missing pixel is reconstructed by peripheral field interpolation. 7. The detection and reconstruction method of dynamic image pixels as claimed in claim 1, wherein the front and back field interpolation method is to average the image information of the former time pixel and the later time pixel of the missing pixel to obtain the Method for reconstructing values of missing pixels. 8. The detection and reconstruction method of dynamic image pixels as claimed in claim 1, wherein the peripheral interpolation method averages the image information of the first position pixel and the second position pixel to obtain the missing pixel reconstruction value method. 9. A method for detecting and reconstructing dynamic image pixels, used for reconstructing missing pixels, the reconstruction steps comprising: Input the time field and multiple reference fields; Scanning out the missing pixel in the time field of the image; Calculate the first difference value of the pixels around the missing pixel, the first difference value is calculated for the surrounding time field having the same scanning state as the time field of the missing pixel; Calculating the second difference value of the surrounding pixels around the missing pixel, the second difference value is calculated for the surrounding time field having a different scanning state from the time field of the missing pixel; Comparing the first difference value, the second difference value with the first critical value and the second critical value; Modify the value of the static counter, and correct it according to the comparison result between the difference value and the critical value; Reconstructing the missing pixel, reconstructing the missing pixel according to the value of the static counter by front-to-back field interpolation or boundary-oriented peripheral field interpolation; output reconstructed de-interlaced pixels; and Calculates whether the scan and rebuild steps are complete. 10. The method for detection and reconstruction of dynamic image pixels as claimed in claim 9, wherein the missing pixels are pixels discarded during sequential scan to interlaced scan conversion. 11. The method for detection and reconstruction of dynamic image pixels according to claim 9, wherein the plurality of reference fields are a plurality of time fields before and after the time field in which the missing pixel is located in the upper field. 12. The method for detection and reconstruction of dynamic image pixels according to claim 9, wherein the plurality of reference fields are a plurality of time fields before and after the time field where the missing pixel is located in the lower field. 13. The method for detecting and reconstructing dynamic image pixels as claimed in claim 9, wherein the difference value is the absolute value of the difference between the previous time field of the missing pixel and the surrounding pixels of the back time field sum. 14. The method for detecting and reconstructing dynamic image pixels as claimed in claim 9, wherein if the first difference value is smaller than the first critical value and the second difference value is smaller than the second critical value, the static counter The value is increased by 1, and if the first difference value is greater than or equal to the first critical value or the second difference value is greater than or equal to the second critical value, the value of the static counter is cleared to 0. 15. The method for detection and reconstruction of dynamic image pixels as claimed in claim 9, wherein if the value of the static counter is greater than or equal to a critical count value, the missing pixel is reconstructed by the front and back field interpolation method, if the static counter If the value of the counter is less than the count threshold, the missing pixel is reconstructed by the border-directed peripheral field interpolation method. 16. The method for detecting and reconstructing dynamic image pixels as claimed in claim 9, wherein the front and back field interpolation method is to average the image information of the former time pixel and the later time pixel of the missing pixel to obtain the Method for reconstructing values of missing pixels. 17. The detection and reconstruction method of dynamic image pixels as claimed in claim 15, wherein if the border-guided peripheral interpolation method selects the function difference of the pixels at the peripheral diagonal positions of the missing pixel with the smallest difference, the diagonal The average value of the position pixels is the reconstructed value of the missing pixel. 18. The method for detecting and reconstructing dynamic image pixels as claimed in claim 17, wherein after the missing pixel is reconstructed by the boundary-guided peripheral field interpolation method, the reconstruction value of the missing pixel is further limited by an intermediate value filtering method error. 19. The detection and reconstruction method of dynamic image pixels as claimed in claim 13, wherein the intermediate value filtering method is to find the intermediate value between the boundary-guided peripheral field interpolation method and the upper and lower relative pixels as the missing pixel reconstruction value method.

Images