Multispectral Image Compression Based on DSC Combined with CCSDS-IDC

Algorithm 1

The HT-transform algorithm for one block.

Input: 4 × 4 or 2 × 2 wavelet coefficients block, denoted as 𝑓_𝑖, 𝑖=1, 2, …, N;

Output: Compressed bit-stream

 foreach block 𝑓_𝑖   do

 {

  Block 𝑓_𝑖 is project on HT base 𝐵_𝑏 to generate the HT-transformed coefficients𝑓_𝑏:

        𝑓_𝑏=^𝑀_𝑚=1𝑓,𝜙_𝑏,𝑚𝜙_𝑏,𝑚;

    The quantizer 𝑄_𝑞 quantize the coefficients 𝑓_𝑏:

            𝑓_𝑏Δ=𝑄_Δ𝑓_𝑏;

    Compute Lagrangian rate-distortion cost function:

        𝐿(𝑓_𝑏Δ)=𝐷(𝑓_𝑏Δ)+𝜆𝑅𝑓_𝑏Δ;

  }

  End

   Save 𝑓^𝑏∗_𝑞, which minimize the Lagrangian rate-distortion cost function 𝐿(𝑓^𝑏_𝑞):

       𝑓^𝑏∗_𝑞=arg_{{𝑓𝑏Δ}1𝑏=0}𝐿(𝑓_𝑏Δ);

  Save identification base index 𝑏^∗.

 }

 End