Convert uint8 vector to float

Hi All,

i have a .mf4-measurement data in uint8:

1×20 uint8 row vector

132 132 132 132 132 132 133 133 133 133 134 138 170 214 222 226 224 216 215 215

is there a way to convert it to these physical values (float)?

0.01 0.01 0.01 0.01 0.01 0.01 0.02 0.02 0.02 0.02 0.03 0.07 0.39 0.83 0.91 0.95 0.93 0.85 0.84 0.84

Thank u and kind regards,

[Les Beckham]

Normally this would be as simple as converting the input data to double and multiplying by a scale factor. Your example values, however don't have a simple linear relationship between the input data and the example output data (see below).

Can you clarify how your example "physical values" were generated?

in = uint8([132   132   132   132   132   132   133   133   133   133   134   138   170   214   222   226   224   216   215   215])

in = 1×20

132 132 132 132 132 132 133 133 133 133 134 138 170 214 222 226 224 216 215 215

out = [0.01 0.01 0.01 0.01 0.01 0.01 0.02 0.02 0.02 0.02 0.03 0.07 0.39 0.83 0.91 0.95 0.93 0.85 0.84 0.84]

out = 1×20

0.0100 0.0100 0.0100 0.0100 0.0100 0.0100 0.0200 0.0200 0.0200 0.0200 0.0300 0.0700 0.3900 0.8300 0.9100 0.9500 0.9300 0.8500 0.8400 0.8400

SF = out./double(in)

SF = 1×20

0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 0.0002 0.0002 0.0002 0.0002 0.0002 0.0005 0.0023 0.0039 0.0041 0.0042 0.0042 0.0039 0.0039 0.0039

plot(SF)
grid on

[hohohihi]

HI Les,

thank u for ur answer.

Thats my problem, i've tried many conversion, also with im2double, but unfortunately it did not work.

I use ETAS-MDA Tool to analyze my .mf4-measurement, on that tool i do have 3 options to view my data:

• Physical Value
• HEX
• and Decimal (RAW)

When i set the dataview to physical value, i got that values (0.01 0.01 0.01 0.01 ...).

When i set the datareview to decimal (RAW), i got the uint8 values (132 132 132 132...)

it happened too, when i imported my mf4-measurement to MATLAB by mdf-reader, thats why i'm trying to find a way to get the physical values by converting..

Kind regards,

Answer 1

[Star Strider]

Since you apparently know the mapping, one option is to simply calculate a linear regression (assuming that a linear relation exists) —

u = uint8([132   132   132   132   132   132   133   133   133   133   134   138   170   214   222   226   224   216   215   215]);
v = [0.01 0.01 0.01 0.01 0.01 0.01 0.02 0.02 0.02 0.02 0.03 0.07 0.39 0.83 0.91 0.95 0.93 0.85 0.84 0.84];

B = [132 1; 226 1] \ [0.01; 0.95]                       % Linear Regression

B = 2×1

0.0100 -1.3100

f = [double(u(:)) ones(size(u(:)))] * B;

figure
plot(double(u), v, '.', 'DisplayName','Data')
hold on
plot(double(u), f, '-', 'DisplayName','Linear Regression')
hold off
grid
xlabel('u')
ylabel('f')
legend('Location','N')

It would likely help to know what the data represent and what the actual conversion is. This approach is a guess that may not work on data other than those in the range of the ‘u’ vector.

.

[hohohihi]

Very good idea, I'll check few of measurement data to confirm the linear relation.

Thank you Star!! :-)

[Star Strider]

As always, my pleasure!