Question

我正在尝试用VHDL计算图像直方图。我有24位强度值，因为强度水平太多，我必须创建箱（最大1920）。我在比较每个输入像素的强度值和决定bin值时遇到麻烦，因此我可以使用bin值作为地址并更新BRAM中的计数。

我在模拟时遇到错误：达到了迭代限制

如果有不同的方法来计算垃圾箱，请告诉我。谢谢！

声明 -

signal data_in : std_logic_vector(23 downto 0);
signal bin : std_logic_vector(11 downto 0);
signal s_data_in : integer ;
signal s_binsize : integer :=2048;
signal s_bin : integer  ;
signal s_bincount : integer :=0 ;
signal s_temp: integer :=2048;

p_clk: process
begin
  s_clk <= '0';
  while (s_SimuDone /= 'H') loop
     wait for C_CLOCK_PERIOD/2;
     s_clk <= not s_clk;
  end loop;
  wait;
end process p_clk;

stimulus: process 
begin
  data_in <= x"0000ff";
  wait for C_CLOCK_PERIOD*10;
  data_in <= x"00000f";
  wait for C_CLOCK_PERIOD*10;
  data_in <= x"000fff";
  wait for C_CLOCK_PERIOD*10;
  data_in <= x"0fffff";
  wait;
end process stimulus;
s_data_in <= to_integer(unsigned (data_in));
bin <= std_logic_vector(to_unsigned(s_bin, 12));

process (s_data_in, s_bincount)
begin
    if (s_data_in < s_binsize) then
        s_bin <= 0;
    else if (s_data_in < s_temp) then
            s_bin <= s_bincount;
            s_bincount <= 0;
            s_temp <= 2048;
         else
            s_bincount <= s_bincount + 1;
            s_temp <= s_temp + s_binsize;
         end if;       
    end if;
end process;

Answer 1

由于语句s_bincount <= s_bincount + 1，错误与一个模拟增量中的过多迭代有关。由于设计是同步的，因此需要指定时钟。但是，为了满足您的需求，您可以将测试平台更改为以下内容：

signal data_in : std_logic_vector(23 downto 0);
signal bin : std_logic_vector(11 downto 0);
signal s_data_in : integer ;
signal s_binsize : integer :=2048;
signal s_bin : integer  ;
signal s_bincount : integer :=0 ;
signal s_temp: integer :=2048;

p_clk: process
begin
  s_clk <= '0';
  while (s_SimuDone /= 'H') loop
     wait for C_CLOCK_PERIOD/2;
     s_clk <= not s_clk;
  end loop;
  wait;
end process p_clk;

stimulus: process 
begin
  data_in <= x"0000ff";
  wait for C_CLOCK_PERIOD*10;
  data_in <= x"00000f";
  wait for C_CLOCK_PERIOD*10;
  data_in <= x"000fff";
  wait for C_CLOCK_PERIOD*10;
  data_in <= x"0fffff";
  wait;
end process stimulus;
s_data_in <= to_integer(unsigned (data_in));

bin <= data_in(22 downto 11);

如果您愿意，可以将 bin 设为时钟变量，但在这种情况下您不需要。 bin 也应该是13位，而不是12位。

请模拟这个并告诉我。

VHDL中的直方图计算

1 个答案: