Hough transformation deinterlacing circuitry

• US 9,338,392 B1
• Filed: 03/28/2014
• Issued: 05/10/2016
• Est. Priority Date: 03/28/2014
• Status: Active Grant

• Current Assignee: Altera Corporation (Intel Corporation)
• Original Assignee: Altera Corporation (Intel Corporation)
• Inventors: Harris, Jon Michael
• Primary Examiner(s): Harold, Jefferey
• Assistant Examiner(s): Haiem, Sean
• Application Number:
  US14/229,480
• Time in Patent Office: 774 Days
• US Class Current: 1/1
• CPC Class Codes: H04N 7/012 Conversion between an interlaced and a progressive signal for graphics images G09G2310/0229
• CPC Class Codes: H04N 7/0142 the interpolation being edge adaptive edge detection in television signals H04N5/142; edge-driven scaling G06T3/403

Abstract

Integrated circuits with interpolating circuitry for deinterlacing video data are provided. Each frame of video data may include a field of pixels from every-other row of the frame. Edge detection circuitry may generate an edge map for the field. The edge map may be filtered to remove weak edges from the map. Cartesian Hough transformation circuitry may perform Cartesian Hough transforms on the filtered edge map to generate a Cartesian Hough space histogram. The Cartesian Hough space histogram may be filtered to remove bins having insufficient counts. Offset mask generation circuitry may generate an offset mask based on the Cartesian Hough space histogram. Interpolation circuitry may generate interpolated pixel values for missing rows of the field using pixels from the field and the offset mask. The interpolating circuitry may generate deinterlaced video data that is free from image artifacts along low-angle edges using the interpolated pixel values.

Claims

1. A method of deinterlacing video data using interpolation circuitry, comprising:

with the interpolation circuitry, generating an edge map that comprises a plurality of edge map values associated with a field of the video data, wherein generating the edge map comprises;

comparing the plurality of edge map values to an edge map threshold value; and

filtering out edge map values that are less than the edge map threshold from the edge map;

with the interpolation circuitry, adjusting the edge map threshold value based on an amount of processing resources available to the interpolation circuitry; and

with the interpolation circuitry, performing a Cartesian Hough transform on the edge map associated with the field of the video data.

2. The method defined in claim 1, wherein the interpolation circuitry comprises Sobel transformation circuitry and wherein generating the edge map comprises:

with the Sobel transformation circuitry, performing a Sobel transform on the field to generate the edge map.

3. The method defined in claim 1, wherein performing the Cartesian Hough transform comprises:

generating a Cartesian Hough space histogram based on the edge map.

4. The method defined in claim 3, further comprising:

with the interpolation circuitry, generating a filtered Cartesian Hough space histogram by comparing the Cartesian Hough space histogram to a Hough space threshold and filtering out bins of the Cartesian Hough space histogram having a bin count that is less than the Hough space threshold.

5. The method defined in claim 3, further comprising:

with the interpolation circuitry, generating an offset mask for the field based on the Cartesian Hough space histogram.

6. The method defined in claim 5, further comprising:

with the interpolation circuitry, generating deinterlaced video data based on the offset mask and the field of the video data.

7. The method defined in claim 6, wherein generating the deinterlaced video data based on the offset mask and the field of the video data comprises:

identifying a plurality of pixels in the field to combine based on a plurality of mask values in the offset mask; and

combining the identified plurality of pixels to generate an interpolated pixel value.

8. The method defined in claim 1, wherein performing the Cartesian Hough transform comprises:

drawing a set of lines through at least one point on the edge map, wherein each line in the set of lines has a respective slope value; and

computing a y-intercept value for each of the lines in the set of lines.

9. The method defined in claim 8, wherein each slope value is less than or equal to one and greater than or equal to negative one.

10. The method defined in claim 8, wherein drawing the set of lines through the at least one point on the edge map comprises:

drawing only low-angle lines having an angle relative to horizontal of less than or equal to 45 degrees.

11. The method defined in claim 8, wherein performing the Cartesian Hough transform further comprises:

drawing a line in Cartesian Hough space using the y-intercept values and the respective slope values; and

accumulating a Cartesian Hough space histogram based on the line in the Cartesian Hough space.

12. Circuitry for deinterlacing an interlaced frame of pixels, comprising:

edge detection circuitry that generates an edge map by performing edge detection operations on a field of the interlaced frame;

Hough transformation circuitry that generates a Hough space accumulation map based on the edge map, the Hough transformation circuitry comprising;

Cartesian Hough space bin accumulation circuitry that accumulates bins in a Cartesian Hough space histogram based on lines in a Cartesian Hough space;

offset mask generation circuitry that generates an offset mask for the field based on the Hough space accumulation map, wherein the Cartesian Hough space bin accumulation circuitry is coupled to the offset mask generation circuitry; and

interpolation circuitry that generates a deinterlaced frame based on pixels in the field and the offset mask.

13. The circuitry defined in claim 12, wherein the Hough transformation circuitry comprises:

linear calculation circuitry coupled to the edge detection circuitry that determines intercept values for a set of lines passing through a point on the edge map.

14. The circuitry defined in claim 13, wherein the linear calculation circuitry determines the intercept values only for lines having an angle relative to horizontal that is less than or equal to 45 degrees and greater than or equal to −

45 degrees.

15. The circuitry defined in claim 13, wherein the Hough transformation circuitry further comprises:

Cartesian Hough space drawing circuitry coupled to the linear calculation circuitry that draws the lines in the Cartesian Hough space based on the intercept values.

16. The circuitry defined in claim 12, further comprising:

register circuitry coupled between the Hough transformation circuitry and the edge detection circuitry; and

edge map comparator circuitry coupled between the register circuitry and the edge detection circuitry that compares the edge map to an edge threshold value.

17. The circuitry defined in claim 12, further comprising:

comparator circuitry coupled between the Hough transformation circuitry and the offset mask generation circuitry that compares the Hough space accumulation map to a Hough space threshold value.

18. The circuitry defined in claim 12, wherein the circuitry is formed on a programmable integrated circuit having programmable logic regions.

19. A method of deinterlacing video data using an integrated circuit, comprising:

with bin accumulation circuitry on the integrated circuit, accumulating a Hough space accumulation map associated with a field of the video data;

with offset mask generation circuitry on the integrated circuit, generating an offset mask based on the accumulated Hough space accumulation map; and

with interpolation circuitry on the integrated circuit, generating an interpolated pixel value at a pixel location in the field by combining pixel values in the field based on the offset mask.

20. The method defined in claim 19, further comprising:

with edge detection circuitry on the integrated circuit, generating a map of edge map values associated with the field of the video data.

21. The method defined in claim 20, further comprising:

with the interpolation circuitry, determining whether the pixel location is on an edge in the field of video data based on the edge map values.

22. The method defined in claim 21, further comprising:

in response to determining that the pixel location is on the edge in the field, generating the interpolated pixel value at the pixel location by combining the pixel values in the field based on the offset mask.