Rocket Auto Recognition And Trace System Based On Digital Signal Processing

At the period of satellite launching, Operator of TV can't track the rocket normally and loses the tracking target, because of the factors such as the greater acceleration of rocket in initial flight stage, the rapid velocity change, the flight direction change of the rocket, technical ability and psychological factor of the operator, etc. Especially when the rocket launching direction changes greatly or the rocket flying deviation appears, it is very difficult to manipulate the camera to track the rocket normally by manual method, which induces the commander can't know the rocket's flight state in season and makes correct judgement. In order to reduce the operator's working intensity and pressure, help the completion of the task better, the dissertation has carried on research about realizing the camera auto-tracking the target and acquiring the rocket flying images.Firstly, the researching actuality in domestic and foreign academia about digital signal disposal technology, auto-target identification technology and digital image disposal algorithm, etc, is introduced.The pivotal algorithms about rocket auto-target identification and tracking are analyzed and compared. The crucial technology of rocket auto-target identification and tracking is the algorithm about image disposal and target tacking. Therefore, all kinds of image predisposal algorithms are analyzed firstly, some image-division algorithms in common use are simulated and compared secondly, including the image-division algorithm based on gray degree gate-limitation, the image-division algorithm based on boundary checking, the tacking algorithm based on moving image sequence, etc. Finally, the target tacking algorithm is analyzed and researched on detailedly. The algorithms research on some crucial images which abstracted from rocket launching kinescope, by simulation the advantages and disadvantages of all algorithms are put forward. Then the suitable algorithm is confirmed for the dissertation. The difference algorithm among frames is ameliorated on basis on fact instance, the simulation result demonstrates that the ameliorated difference algorithm is valid and favorable.The general hardware composing framework of the identification system is put forward, the idiographic realization method of rocket auto-target identification system is expounded. By comparing the characteristics of different CMOS chips, the crucial chips are confirmed. Then the principle charts of the main hardware modules are introduced according to the confirmed chips. The working principle and design method of all modules are demonstrated.Finally, the pivotal algorithm of software for the rocket auto-target identification system is expatiated, the flow structure of the algorithm is presented, and the method of exploiting DSP through C language programming and the admixture programming of C language and assembly language is described.