A TBD algorithm based on improved Randomized Hough Transform for dim target detection

For the detection of dim target, the track-before-detect(TBD) methodology has been employed to integrate the target's energy through a time sequence of frames. A TBD algorithm based on improved Randomized Hough Transform for dim target detection is proposed in this paper. This algorithm uses the sequence number of frames to make sure that the point pairs are selected randomly from different frames, avoiding the unreasonable situation that initiates target's track in the same frame data. Second, it also introduces a new vote method. Based on the minimum Euclidean distance criterion, this vote method finds the optimal parameter cell to vote, making the vote result better than traditional Randomized Hough Transform (RHT). In addition, we not only increase the poll of the optimal parameter cell but also update the corresponding parameter, avoiding the large deviation between the recovered track and the real track. By simulation, we conclude that this algorithm can detect dim target more speedily and accurately than traditional RHT, especially under the background of low Signal to Noise Ratio (SNR).