olsen wrote:
hi thanks to IOhannes indications i'm feeding my pix_mulitblob output into the iemmatrix now. as long as there're enough blobs in the picture it works pretty nice. but as soon as there're less blobs than given in the argument to pix_multiblob i get: error: $43: argument number out of range error: $44: argument number out of range
without looking at the patch, i can tell you that this is because you are actually _not_ using iemmatrix for handling the matrix. instead your patch tries to manually parse the data using dollargs and whatelse; which obviously fails if there are too few arguments (because the matrix is too small)
... for all the arguments of undetected blobs. and as soon as there's no blob in the picture the pd-shell reports: matrix : corrupt matrix passed
this indicates that you _are_ using iemmatrix here. if no blobs are detected, pix_multiblob will output a somewhat illegal matrix with zero rows (and a number of columns).
now iemmatrix complains that this is not a legal matrix. probably putting a [mtx_check] helps to get rid of the error (or not)
do i do something wrong?
no.
fgmasdr IOhannes