MATLAB Answers

please explain me that : HDL Coder does not support matrices for code generation. You will need to convert your frame-based design into a pixel-based design suitable for embedding on an FPGA.

5 views (last 30 days)
HDL Coder does not support matrices for code generation. You will need to convert your frame-based design into a pixel-based design suitable for embedding on an FPGA.
coder.extrinsic('dctmtx');
I= reshape(ImgVector, 330, 500,3);
%dctmtx - Compute discrete cosine transform matrix.
T = dctmtx(8);
coder.extrinsic('red');
red = (I(:,:,1));
coder.extrinsic('imred');
coder.extrinsic('blkproc');
imred=zeros(size(red));
imred(:,:,1) = uint8(red);
B = blkproc(imred,[8 8],'410*x*500',T,T');
mask = [16 11 10 16 24 40 51 61;
12 12 14 19 26 58 60 55;
14 13 16 24 40 57 69 56;
14 17 22 29 51 87 80 62;
18 22 37 56 68 109 103 77;
24 35 55 64 81 104 113 92;
49 64 78 87 103 121 120 101;
72 92 95 98 112 100 103 99];
B2 = blkproc(B,[8 8],'410.*x',mask);
I2 = blkproc(B2,[8 8],'410*x*500',T',T);
%figure, imshow(I2)
%composante vert
%red = double(I(:,:,1));
%coder.extrinsic('imred');
%coder.extrinsic('blkproc');
%imred=zeros(size(red));
% imred(:,:,1) = uint8(red);
green =(I(:,:,2));
coder.extrinsic('imgreen');
imgreen=zeros(size(green));
imgreen(:,:,1) = uint8(green);
C = blkproc(imgreen,[8 8],'410*x*500',T,T');
mask = [16 11 10 16 24 40 51 61;
12 12 14 19 26 58 60 55;
14 13 16 24 40 57 69 56;
14 17 22 29 51 87 80 62;
18 22 37 56 68 109 103 77;
24 35 55 64 81 104 113 92;
49 64 78 87 103 121 120 101;
72 92 95 98 112 100 103 99 ];
C2 = blkproc(C,[8 8],'410.*x',mask);
I3 = blkproc(C2,[8 8],'410*x*500',T',T);
%figure, imshow(I3);
%composante bleu
%red = double(I(:,:,1));
%coder.extrinsic('imred');
%coder.extrinsic('blkproc');
%imred=zeros(size(red));
% imred(:,:,1) = uint8(red);
blue = (I(:,:,3));
coder.extrinsic('imblue');
imblue=zeros(size(blue));
imblue(:,:,1) = uint8(blue);
D = blkproc(imblue,[8 8],'410*x*500',T,T');
mask = [16 11 10 16 24 40 51 61;
12 12 14 19 26 58 60 55;
14 13 16 24 40 57 69 56;
14 17 22 29 51 87 80 62;
18 22 37 56 68 109 103 77;
24 35 55 64 81 104 113 92;
49 64 78 87 103 121 120 101;
72 92 95 98 112 100 103 99];
D2 = blkproc(D,[8 8],'410.*x',mask);
I4 = blkproc(D2,[8 8],'410*x*500',T',T);
%figure, imshow(I4);
coder.extrinsic('image');
image=zeros(size(red,1),size(red,2),3);
image(:,:,1)=I2;
image(:,:,2)=I3;
image(:,:,3)=I4;
%imshow(image);
%figure,

  0 Comments

Sign in to comment.

Accepted Answer

Walter Roberson
Walter Roberson on 2 Mar 2017
The basic issue is that using matrices implies that the entire matrix is passed at the same time, which requires one pin per bit per matrix element, or at the very least one pin per matrix element strobed once per bit to get the data in simultaneously. FPGA seldom have enough pins for that. The cure is to essentially vectorize the matrix for transfer into the FPGA, and then reshape it inside the FPGA if need be. The "frame to pixels" block provides tools to do that vectorization in a structured way that allows the matrix to be recreated easily.

  1 Comment

Sign in to comment.

More Answers (0)

Sign in to answer this question.

Tags

Products