Count and stats of blobs differs between dat and bru

[jiho]

While investigating #19 , I read the .bru and .dat files for
pproj 33 = uvp5_sn002zd_cascade2011
psampleid 11531 = cx01_14, raw = HDR20110313090333

I recomputed the number of blobs and their mean area and mean meangrey from the bru and compared those with the starts in the dat. This table shows the content of the dat + the last 3 columns which are my recomputation from the bru.

    index              image press nb blobs P-G mean area P-G mean grey P-G nb blobs G mean grey G n_blob   m_area   m_grey
 1:    25 20110313090335_359  16.4          123             4            30          0           0     88 5.454545 30.98864
 2:    26 20110313090335_453  16.4          540             5            31          0           0    382 5.945026 33.53141
 3:    27 20110313090335_562  16.4          565             5            31          0           0    397 6.138539 33.23174
 4:    28 20110313090335_671  16.4          569             5            31          0           0    396 5.997475 33.60101
 5:    29 20110313090335_781  16.4          602             4            31          0           0    399 6.002506 34.03008
 6:    30 20110313090335_875  16.4          601             5            31          0           0    431 6.041763 33.22738
 7:    31 20110313090335_984  16.3          610             5            31          0           0    419 6.116945 33.25298
 8:    32 20110313090336_093  16.3          580             5            31          0           0    394 6.081218 33.50761

Is it expected that the number of blobs is higher in the dat? In the bru, the smallest area is 4 pixels. Can the difference come from the fact that the info for the smallest objects (4 pixels) is not stored in the bru?