RFI Mitigation

A general VDIF stage for RFI removal currently exists in kotekan.

VDIF

VDIF RFI removal is implemented in 3 different manners.

1. During the read

The Kotekan Stage nDiskFileRead can be configured to do VDIF rfi detection/zeroing as it reads.

To configure nDiskFileRead for rfi removal, example config:

replay:
  kotekan_stage: nDiskFileRead
  num_disks: 10
  disk_base: /mnt/
  disk_set: A
  out_buf: vdif_input_buf_0
  capture: 20170426T110023Z_ARO_raw
  sk_step: 16384 #(The time cadence for kurtosis measurements, in units of time samples)
  rfi: True #(RFI ON/OFF)
  rfi_sensitivity: 3 #(The sensitivity of the kurtosis threshold, lower is more sensitive)
  normalize: True #(Whether or not to normalize before kurtosis measurment)

Relevant files:

• /ch_gpu/lib/processes/nDiskFileRead.cpp (.hpp)

2. During the power integration

The Kotekan Stage computeDualpolPower can be configured to do VDIF rfi detection/zeroing as it integrates.

To configure computeDualpolPower for rfi removal, example config:

power:
  kotekan_stage: computeDualpolPower
  rfi_removal: True #(RFI ON/OFF)
  rfi_sensitivity: 3 #(The sensitivity of the kurtosis threshold, lower is more sensitive)
  rfi_backfill: False #(Dumb Backfill of zeroed data, not recommended for use unless you need a pretty picture)
  vdif_in_buf: gpu_output_buffer_0
  power_out_buf: output_power_buf

Relevant files:

/ch_gpu/lib/processes/computeDualpolPower.cpp (.hpp)