Recognition of Weak Infrared Target Based on DTCNN in the Complex Background

Infrared images are widely used for many purposes. A recognition system of IR targets is essential for equipments that have to work in dark. In this paper, we proposed an IR image weak target recognition algorithm by discrete-time cellular neural network (DTCNN). We constituted an IR image weak target recognition system with DTCNN. The proposed method only requires a 3×3 neighborhood. We used a variable as the bias parameter, which is defined by the position and the value of pixels. Due to the alterable bias parameter, the proposed method can effectively remove the influence of noise. Our system contains "Preprocessing", "Feature extraction", and "Recognition", and we have adopted DTCNN in "Feature extraction" part. For simplicity, we used a normalization algorithm as a preprocessing method. We demonstrated a usefulness of the proposed algorithm in recognition experiments. Three IR images are used. As a result, the good recognition results were obtained.