KEYWORDS: Mining, Land mines, Data modeling, Sensors, Expectation maximization algorithms, General packet radio service, Ground penetrating radar, Metals, Detection and tracking algorithms, Feature extraction
We propose a landmine detection algorithm that uses ensemble discrete hidden Markov models with context
dependent training schemes. 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 is based on clustering in the log-likelihood space.
First, one HMM is fit to each of the N individual sequence. For each fitted model, we evaluate the log-likelihood
of each sequence. This will result in an N x N log-likelihood distance matrix that will be partitioned into K
groups. In the second step, we learn the parameters of one discrete HMM per group. We propose using and
optimizing various training approaches for the different K groups depending on their size and homogeneity. In
particular, we will investigate the maximum likelihood, and the MCE-based discriminative training approaches.
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.
KEYWORDS: Mining, Data modeling, Land mines, Sensors, General packet radio service, Feature extraction, Ground penetrating radar, Detection and tracking algorithms, Artificial neural networks, Metals
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.
KEYWORDS: Mining, Land mines, Data modeling, Ground penetrating radar, General packet radio service, Sensors, Wavelets, Feature extraction, Analytical research, Metals
In this paper, we propose an efficient Discrete Hidden Markov Models (DHMM) for landmine detection that rely
on training data to learn the relevant features that characterize different signatures (mines and non-mines), and
can adapt to different environments and different radar characteristics. Our work is motivated by the fact that
mines and clutter objects have different characteristics depending on the mine type, soil and weather conditions,
and burial depth. Thus, ideally different sets of specialized features may be needed to achieve high detection and
low false alarm rates. The proposed approach includes three main components: feature extraction, clustering,
and DHMM. First, since we do not assume that the relevant features for the different signatures are known a
priori, we proceed by extracting several sets of features for each signature. Then, we apply a clustering and
feature discrimination algorithm to the training data to quantize it into a set of symbols and learn feature
relevance weights for each symbol. These symbols and their weights are then used in a DHMM framework to
learn the parameters of the mine and the background models. Preliminary results on large and diverse ground
penetrating radar data show that the proposed method outperforms the basic DHMM where all the features are treated equally important.
KEYWORDS: Mining, Land mines, General packet radio service, Wavelets, Sensors, Metals, Data modeling, Prototyping, Ground penetrating radar, Feature extraction
We propose a general method for detecting landmine signatures in vehicle mounted ground penetrating radar (GPR) using
discrete hidden Markov models and Gabor wavelet features. Observation vectors are constructed based on the expansion
of the signature's B-scan using a bank of scale and orientation selective Gabor filters. This expansion provides localized
frequency description that gets encoded in the observation sequence. These observations do not impose an explicit structure
on the mine model, and are used to naturally model the time-varying signatures produced by the interaction of the GPR
and the landmines as the vehicle moves. The proposed method is evaluated on real data collected by a GPR mounted on
a moving vehicle at three different geographical locations that include several lanes. The model parameters are optimized
using the BaumWelch algorithm, and lane-based cross-validation, in which each mine lane is in turn treated as a test set
with the rest of the lanes used for training, is used to train and test the model. Preliminary results show that observations
encoded with Gabor wavelet features perform better than observation encoded with gradient-based edge features.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.