Pulse Coupled Neural Network Algorithm for Object Detection in Infrared Image

This paper describes an object detection algorithm based on pulse coupled neural networks (PCNN). To implement real time fusion of infrared surveillance object with an augmented background, the objects must be detected effectively, but the precision are not required. The simple frame difference method is adopted. We use an improved pulse coupled neural network to segment the frame difference image. The iteration times can be automatically estimated. This algorithm does not need model training and parameter learning, and it is applied to segment images with noise effectively. Object detection on several types of image is implemented with the proposed method and the experimental results demonstrate its validity.