Carrier-based Aircraft Detection on Flight Deck of Aircraft Carrier with Simulated 3-D Model by Deep Neural Network

Military aircraft detection is always a challenging problem in the development of defence technology since World War II (WWII). The target is usually captured by electromagnetic wave reflection, which is accomplished by radar and CCD camera. The existed studies in this field mainly focused on aircraft detection in airbase or airports so far. It is noteworthy that the aircraft carrier is also an essential vector for the military aircraft, which is more challenging for aircraft detection, especially in the complex sea condition. This paper proposed a novel method of carrier-based aircraft detection on the flight deck of an aircraft carrier with simulated 3-D model. We construct a Parallel Convolutional Neural Network (PCNN) to train the simulated 3-D model dataset, which is also used for the final detection process. Finally, we apply the generalized deep convolutional neural network to detect the real reconnaissance images of the aircraft carrier with carrier-based aircraft on the flight deck. Experiments result show that the proposed method achieves an average detection accuracy of 99.92% in the real reconnaissance images, which contributes to the surveillance and early warning of the actual naval warfare.