We propose a landmine detection algorithm that uses a mixture of discrete hidden Markov models. We hypothesize
that the data are generated by K models. These different models reflect the fact that mines and
clutter objects have different characteristics depending on the mine type, soil and weather conditions, and burial
depth. Model identification could be achieved through clustering in the parameters space or in the feature space.
However, this approach is inappropriate as it is not trivial to define a meaningful distance metric for model
parameters or sequence comparison. Our proposed approach is based on clustering in the log-likelihood space,
and has two main steps. First, one HMM is fit to each of the R individual sequence. For each fitted model, we
evaluate the log-likelihood of each sequence. This will result in an R×R log-likelihood distance matrix that will
be partitioned into K groups using a hierarchical clustering algorithm. In the second step, we pool the sequences,
according to which cluster they belong, into K groups, and we fit one HMM to each group. The mixture of these
K HMMs would be used to build a descriptive model of the data. An artificial neural networks is then used to
fuse the output of the K models. Results on large and diverse Ground Penetrating Radar data collections show
that the proposed method can identify meaningful and coherent HMM models that describe different properties
of the data. Each HMM models a group of alarm signatures that share common attributes such as clutter, mine
type, and burial depth. Our initial experiments have also indicated that the proposed mixture model outperform
the baseline HMM that uses one model for the mine and one model for the background.