A motion tracking of moving object for the ATR system

We propose a motion tracking algorithm based on fuzzy inference and wavelet transform. Wavelet transform (WT) is first used for preprocessing, where the role of WT is reducing the search area and is increasing the processing time. After performing the 3-level WT, a set of WT coefficients which is useful to estimate the motion are calculated. Next, an effective fuzzy inference rule is devised to map numerical WT data into linguistic terms. Based on the proposed fuzzy inference algorithm, we also optimized the similarity of motion vector among successive frames. The experimental results show that the proposed fuzzy rule-based motion tracking scheme has a higher average tracking ratio and can be adaptive to the condition of various circumstances.