You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Currently we are using 3x3 AR barcodes. These have some limitations so we should probably change to a different type.
Here is a chart of barcode types (from artoolkitx):
Matrix code type
Maximum number of markers
Number of bit errors that can be detected
Number of bit errors that can be corrected
AR_MATRIX_CODE_5x5_BCH_22_7_7
127
6
3
AR_MATRIX_CODE_4x4_BCH_13_5_5
32
4
2
AR_MATRIX_CODE_5x5_22_12_5
4095
4
2
AR_MATRIX_CODE_3x3_HAMMING63
8
2
1
AR_MATRIX_CODE_4x4_BCH_13_9_3
512
2
1
AR_MATRIX_CODE_3x3_PARITY65
32
1
0
AR_MATRIX_CODE_3x3
64
0
0
AR_MATRIX_CODE_4x4
8192
0
0
AR_MATRIX_CODE_5x5
4194304
0
0
You can see from this chart that the type we are currently using, AR_MATRIX_CODE_3x3, only allows up to 64 markers. Since we are using 1-16 for flowers and reserving 0, that leaves only 47 markers left for nests, which for the majority of cases would be enough but could be an issue in particularly large environments. Aside from that the 3x3 barcodes without any parity we are using have no error correction and thus a hamming distance of 0. This means there is a higher chance of a marker being misrecognized as a different ID.
I propose we use the AR_MATRIX_CODE_5x5_BCH_22_7_7 type as it provides 127 markers, which would give us up to 111 nests and it has the greatest hamming distance between markers and ability to correct for errors in reading. This seems to be the most recommended one in general as well.
The text was updated successfully, but these errors were encountered:
Currently we are using 3x3 AR barcodes. These have some limitations so we should probably change to a different type.
Here is a chart of barcode types (from artoolkitx):
You can see from this chart that the type we are currently using,
AR_MATRIX_CODE_3x3
, only allows up to 64 markers. Since we are using 1-16 for flowers and reserving 0, that leaves only 47 markers left for nests, which for the majority of cases would be enough but could be an issue in particularly large environments. Aside from that the 3x3 barcodes without any parity we are using have no error correction and thus a hamming distance of 0. This means there is a higher chance of a marker being misrecognized as a different ID.I propose we use the
AR_MATRIX_CODE_5x5_BCH_22_7_7
type as it provides 127 markers, which would give us up to 111 nests and it has the greatest hamming distance between markers and ability to correct for errors in reading. This seems to be the most recommended one in general as well.The text was updated successfully, but these errors were encountered: