Image sensor data fusion using factorized Kalman filter

This paper presents image sensor data fusion strategy using factorized Kalman filter algorithm which has wide range of aerospace applications. This involves locating the target from the images obtained from the two sensors using Centroid tracking Factorized Kalman filter and then fusing the sensor data to get much better information of the target position and velocity. Factorized Kalman filter or UD filter (UDF) is used for predicting the upcoming position and other variables of the target. Fusion is used to reduce the error that occurs due to clutters in image data taken from sensors. Performance of two fusion algorithms that is measurement or data level fusion and state vector fusion are carried out and good results are obtained regarding the position and velocity estimation of the target. Image sensor data fusion (ISDF) is realized using MATLAB tool. The sensor images are synthesized and added with different noise levels in order to represent sensor data obtained in the presence of different atmospheric clutter. Segmentation process and nearest neighbor technique is used to extract the target details from the sensor images.