The recognition of landed aircrafts based on PCNN model and affine moment invariants

• Propose a PCNN (Pulse Coupled Neural Network) + AMI method for aircraft recognition.
• Use the PCNN model and affine moment invariants to form features vector.
• The PCNN model is used to obtain a binary image sequence.
• PCNN + AMI has simple computation, processing step reduction and high recognition ratio.

Currently the most research done on the recognition of landed aircrafts based on its shape feature lie in two cases: either the research has focused on different moment invariants algorithms, only using some already processed binary images to verify results, or the research refers to some integrated recognition system which has multi-step preprocessing. The better solution must consider preprocessing, simple computation and step reduction altogether when the high recognition ratio is guaranteed. For this better solution, we propose a landed aircrafts recognition method based on the PCNN model and affine moment invariants (PCNN+AMI). Our method uses the PCNN model to generate binary sequence of grey intensity images, then it extracts affine moment invariants from the sequence to compose features vector. The experimental results illustrate that our method cannot only get a good ability of anti-geometrical distortion but can also have simple computation and step reduction.