QRDecoders

QRInfo

A struct to store the information of a QR code.

Fields

• version::Int: version of the QR code
• eclevel::Int: error correction level of the QR code
• mask::Int: mask of the QR code
• mode::Int: mode of the QR code
• message::String: decoded message

qrdecode(mat::AbstractMatrix
        ; noerror::Bool=false
        , preferutf8::Bool=true
        , alg::ReedSolomonAlgorithm=Euclidean()
        )::QRInfo

QR code decoder.

If noerror is true, the decoder will raise an Exception(ReedSolomonError/InfoError) when the QR code mat needs error correction.

If preferutf8 is true, the decoder will try to decode the message by UTF8 mode when dealing with Byte mode.

The error correction algorithm is specified by the variable alg(default: Euclidean).

qrdecode(path::AbstractString; keywords...)

QR code decoder.

For more information of the keywords, see qrdecode(mat::AbstractMatrix; keywords...).

qrdecodes(path::AbstractString; keywords...)

QR code decoder for animated QR code.

qrdecompose(mat::AbstractMatrix, noerror=false)

Decompose the QR-Code into its constituent parts.

getqrmatrix(imgpath::AbstractString)

Get the QR-code matrix from an image.

Note that the input must be a standard QR-code image with or without white border, i.e. the image should not contain any non-QR-Code information.

getqrmatrix(mat::AbstractMatrix)

Get the standard QR-code matrix from an image-matrix.

getqrmatrices(imgpath::AbstractString)

Get the standard QR-code matrices from a gif file.

qrdecode_version(version::Int)

Decode version information.(Interger to Integer)

qrdecode_version(mat::AbstractMatrix; noerror=false)

Return the version of the QR-Code.(Matrix to Integer)

qrdecode_format(fmt::Int)::Int

Decode format information.

qrdecode_format(mat::AbstractMatrix; noerror=false)

Return the format of the QR-Code(ErrCorrLevel + mask).

extract_databits(mat::AbstractMatrix, datapos::AbstractMatrix)

Extract data bits from the QR-Code.(Inverse procedure of placedata!)

deinterleave(bytes::AbstractVector, ec::ErrCorrLevel, version::Int)

De-interleave the message, i.e. sperate msgblocks and ecblocks from data bits.

block2bits(blocks::AbstractVector)

Convert a block to bits.

correct_message(msgblock::AbstractVector
               , ecblock::AbstractVector
               , alg::ReedSolomonAlgorithm
               ; noerror::Bool=false)

Error correction of the message block using the given algorithm.

Throw DecodeError if the value noerror is true and the message need error correction.

decodemode(bits::AbstractVector)

Decode mode from the bits of length 4. Note: the Byte mode and the UTF8 mode use the same mode indicator(0100).

decodedata(bits::AbstractVector, msglen::Int, ::Mode)

Decode message from bits without checking the pad_bits.

ReedSolomonAlgorithm

An abstract type for error correction algorithm of Reed Solomon code.

Euclidean <: ReedSolomonAlgorithm

Euclidean algorithm for error correction.

BerlekampMassey <: ReedSolomonAlgorithm

Berlekamp-Massey algorithm for error correction.

RSdecoder(received::Poly, nsym::Int, ::ReedSolomonAlgorithm)

Decode the message polynomial using the given Reed-Solomon algorithm.

berlekamp_massey_decoder(received::Poly, erasures::AbstractVector, nsym::Int)

Berlekamp-Massey algorithm, decode message polynomial from received polynomial(given erasures).

berlekamp_massey_decoder(received::Poly, nsym::Int)

Berlekamp-Massey algorithm, decode message polynomial from received polynomial(without erasures).

euclidean_decoder(received::Poly, erasures::AbstractVector, nsym::Int)

Decode the received polynomial using the Euclidean algorithm(with erasures).

euclidean_decoder(received::Poly, erasures::AbstractVector, nsym::Int)

Decode the received polynomial using the Euclidean algorithm(without erasures).

syndrome_polynomial(received::Poly, nsym::Integer)

Computes the syndrome polynomial S(x) for the received polynomial where nsym is the number of syndromes.

evaluator_polynomial(sydpoly::Poly, errlocpoly::Poly, nsym::Int)

Return the evaluator polynomial Ω(x) where Ω(x)≡S(x)Λ(x) mod xⁿ.

erratalocator_polynomial(::Type{T}, errpos::AbstractVector)

Compute the erasures/error locator polynomial Λ(x) from the erasures/errors positions.

erratalocator_polynomial(sydpoly::Poly, nsym::Int; check=false)

Compute the error locator polynomial Λ(x)(without erasures). The check tag ensures that Λx can be decomposed into products of one degree polynomials.

erratalocator_polynomial(sydpoly::Poly, erasures::AbstractVector, n::Int)

Berlekamp-Massey algorithm, compute the error locator polynomial Λ(x)(given the erased positions). The check tag ensures that Λx can be decomposed into products of one degree polynomials.

forney_algorithm(Λx::Poly, Ωx::Poly, errpos::AbstractVector)

Forney algorithm, returns the error-corrected values. eₖ = 2^{iₖ}⋅Ω(2^{-iₖ}) / Λ'(2^{-iₖ})

haserrors(received::Poly, nsym::Int)

Returns true if the received polynomial has errors. (may go undetected when the number of errors exceeds n)

fillerasures(received::Poly, errpos::AbstractVector, nsym::Int)

Forney algorithm, computes the values (error magnitude) to correct the input message.

Warnning: The output polynomial might be incorrect if errpos is incomplete.

Sugiyama_euclidean_divide(r₁::Poly, r₂::Poly, upperdeg::Int)

Yasuo Sugiyama's adaptation of the Extended Euclidean algorithm. Find u(x), v(x) and r(x) s.t. r(x) = u(x)r₁(x) + v(x)r₂(x) where r(x) = gcd(r₁(x), r₂(x)) or deg(r(x)) ≤ upperdeg.

polynomial_eval(p::Poly, x::Integer)

Evaluates the polynomial p(x) at x in GF256.

derivative_polynomial(p::Poly)

Computes the derivative of the polynomial p.

findroots(p::Poly)

Computes the roots of the polynomial p using Horner's method.

The output will be an empty list if p(x) contains duplicate roots or roots not in GF(256).

reducebyHorner(p::Poly, a::Int)

Horner's rule, find the polynomial q(x) such that p(x)-(x-a)q(x) is a constant polynomial.

getpositions(Λx::Poly)

Caculate positions of errors from the error locator polynomial Λx. Note that the indexs start from 0.

extended_euclidean_divide(r₁::Poly, r₂::Poly)

Return polynomials u(x) and v(x) such that u(x)r₁(x) + v(x)r₂(x) = gcd(r₁(x), r₂(x)).

illustration

Let

\[r_k = u_kr_1 + v_kr_2,\quad k \geq 2\]

Then

\[\begin{aligned} r_0 &= q_0r_1 + r_2 \quad\Rightarrow r_2 = r_0 - q_0r_1,\ where\ r_0=r_2,\ q_0=0\\ r_1 &= q_1r_2 + r_3 \quad\Rightarrow r_3 = r_1 - q_1r_2\\ r_2 &= q_2r_3 + r_4 \quad\Rightarrow r_4 = r_2 - q_2r_3\\ &\vdots\\ r_k &= q_kr_{k+1} + r_{k+2}\Rightarrow r_{k+2} = r_k - q_kr_{k+1}\\ &\phantom{= q_kr_{k+1} + r_{k+2}\Rightarrow r_{k+2}} = u_kr_1 + v_kr_2 - q_k(u_{k+1}r_1 + v_{k+1}r_2)\\ &\phantom{= q_kr_{k+1} + r_{k+2}\Rightarrow r_{k+2}} = (u_k - q_ku_{k+1})r_1 + (v_k - q_kv_{k+1})r_2 \end{aligned}\]

Loop until $r_t = q_tr_{t+1} + 0$, then $r_{t+1}$ is the greatest common factor of $r_1$ and $r_2$.

Here we obtain the recursive formula of $u_k$ and $v_k$.

\[\begin{aligned} u_2, v_2 &= 0, 1,\quad u_3, v_3 = 1, -q_1\\ u_{k+1} &= u_k - q_ku_{k+1},\quad v_{k+1} = v_k - q_kv_{k+1} \end{aligned}\]

InfoError <: Exception

The non-data part of QR-matrix contains error.

For example, Finder pattern, Alignment pattern, Timing pattern, Format information, Version information, matrix size and etc.

DecodeError <: Exception

The data part of QR-matrix contains error.

ReedSolomonError <: Exception

An error occurs during error-correction.