A modularized reconfigurable system for target recognition with multi-DSP processing is designed to reconfigure the target recognition modules and update the distributed target feature libraries through the serial channel to adapt to the varied application. The system is separated into three independent modules and two work modes running at different time slides based on project switch. The modularized reconfiguration module is designed as a minimum security kernel separated from the target recognition module to decrease their coupling and interrelationship. This kind of multi-project design based on cyclic redundancy check presents a more independent and reliable target recognition system with modularized reconfiguration ability.
A stable imaging tracking method based on learning online for ground moving target with multi-DSP processing is
presented in this paper. Background window is set to track and predict the background image and supervise the intruder.
The target learning online based on background prediction revises the accumulated tracking error. Different tracking
strategy during different tracking states and risk level of intruder improves the stability and accuracy of tracking system
especially in a long time of continual tracking. The parallel processing based on multiple DSP makes a real-time tracking
system be possible.
An adaptive tracking method for ground target in FLIR image including centric and eccentric tracking is presented in this
paper. The eccentric tracking is adopted to assist ATR or manual target acquisition by stabilizing the light axis of the
seeker. The tracking adaptability detection at the start of centric tracking avoids tracking bad locked point and decreases
the shift, slide and jump in the subsequent tracking. Combination of periodic template-updating based on scale variance
rate and interruptive template-updating based on supervision of tracking point enhances the adaptability of tracking
template. Supervision, modification of current tracking point and the combination of direct and indirect tracking increase
the stability of tracking system. The multi-scale and size-variable template is adopted to fit the variable target scale. The
detection and adoption of occlusion mask increase the accuracy of the tracking system. The hardware system based on
multi-DSP takes a real-time parallel processing. The trial results show that this method is efficient to reduce the
possibility of shift, slide, and jump in ground target tracking comparing with the traditional correlation tracking method.
A precise tracking algorithm for small target based on event supervision is introduced in this paper. The target chains and
object aggregation are established firstly, Tri-level scan filter contains grey intension filter, shape filter and location filter,
is adopted to implement data relevancy between target chains and object aggregation from coarse to fine. Movement
trend of object to tracking target which is classified into follow, approach and leaving, and events include envelop and
combination are detected, supervised and processed. On the other hand, based on the analysis of the error model for
target centroid estimation, a recursive approach method for target centroid calculation with high precision is adopted in
the algorithm. It combines tracking, recognition and prediction effectively base on the tracking theory of human eyes.
Experimental results show that the method is feasible and effective.