Gamma-ray spectrum classification

1. A gamma-ray spectrum classification apparatus, comprising circuitry configured:

to provide a denoising autoencoder to receive deconvolved gamma-ray spectrum data representing a gamma-ray spectrum of a material to be classified and to denoise the deconvolved gamma-ray spectrum data to determine feature data indicative of one or more features representative of the gamma-ray spectrum data; and

to provide a classification neural network to receive the feature data and to classify the material to be classified as one of a plurality of predetermined classifications using the feature data.

2. The apparatus according to claim 1 , wherein the circuitry is configured:

to provide denoising training data as a target output of the denoising autoencoder, the denoising training data representing a training gamma-ray spectrum for training the denoising autoencoder;

to resample the denoising training data and provide the resampled denoising training data as an input of the denoising autoencoder;

to determine weights of the denoising autoencoder based on the input and the target output; and

to determine the feature data using the determined weights of the denoising autoencoder.

3. The apparatus according to claim 2 , wherein the training gamma-ray spectrum comprises a background gamma-ray spectrum combined with a predetermined gamma-ray spectrum of a sample of material.

4. The apparatus according to claim 3 , wherein the circuitry is configured to receive background gamma-ray spectrum data representing the background gamma-ray spectrum from a gamma-ray detector located at a location at which the material to be classified is to be located.

5. The apparatus according to claim 1 , wherein the circuitry is configured:

to receive classification training data representing one or more training gamma-ray spectra for training the classification neural network;

to receive classification information classifying each of the one or more training gamma-ray spectra, the classification information indicating one of the plurality of predetermined material classifications for each of the one or more training gamma-ray spectra;

to determine weights of the classification neural network based on the classification training data and the classification information;

to classify the material to be classified using the determined weights of the classification neural network.

6. The apparatus according to claim 1 , wherein the material to be classified is classified as either a naturally occurring radiation material (NORM) or a non-naturally occurring radiation material (non-NORM).

7. The apparatus according to claim 6 , wherein, when the material to be classified is classified as a NORM material, the material is further classified as one of a plurality of predetermined types of NORM material.

8. A radiation portal monitor comprising a gamma-ray spectrum classification apparatus according to claim 1 .

9. A gamma-ray spectrum classification method, comprising:

receiving deconvolved gamma-ray spectrum data representing a gamma-ray spectrum of a material to be classified and denoising the deconvolved gamma-ray spectrum data to determine feature data indicative of one or more features representative of the gamma-ray spectrum data using a denoising autoencoder; and

using the feature data to classify the material to be classified as one of a plurality of predetermined classifications using a classification neural network.

10. A non-transitory machine-readable medium comprising a program for controlling a computer to perform a method according to claim 9 .

11. A storage medium storing a computer program according to claim 10 .