Research On Image Mosaicing Technique In Automobile Chassis Security Visual Inspection System

In recent years, the car is more and more popular. The criminal carries dangerousand prohibited items using motor vehicle, which has become troubling seriousproblem all over the world. The terrorist organization is so rampant that theautomobile explosion unceasingly happened all over the world. In order to safeguardthe society be stable, make the lives and the possessions be safe, all of the countriesreinforce the security monitoring measure. The terrorists often hide the explosion inthe bottom of the automobile to avoid being in sight. Also, some people are hidden inthe bottom of the automobile to enter some country illegally. Therefore, the security ofautomobile chassis has become the focus of the security inspection.An automobile chassis security visual inspection system is designed and realizedin this paper. In the thesis, we firstly expounded the design of the whole systemsolutions, analyzed the working principle of the system, and then introduced the mainfunction modules and the key technology method. According to the dynamic vehicleimaging characteristics, we used the line scanning technology to achieve the images ofvehicle bottom and license plate, and then transmitted these images to the remoteserver using the network. According to the requirements of the system, the imagesmosaics technology is studied in the paper. The proposed image registration algorithmcombined fractal signature and gray correlation. Firstly, we found the regions thatmaybe the same or similarly between two images using the fractal signature in thecoarse registration. Then, we eliminated the false matching regions by use of graycorrelation, and gain the final matching result. The experiment shows that thealgorithm can enhance the registration speed and ensure the stability of theregistration. We proposed a new image fusion algorithm based on Curvelet transform.The experimental results show that the fusion images have higher information relateddegree and less difference with original image. The new algorithm has effectivelyimproved the quality of image mosaics.Finally, we designed the client software and server software of the system, andsolved the problems in the actual testing. The various performance of the system workwell. Under the condition of guaranteeing reliability, the system can greatly increasethe bottom inspection’s automation and will be widely used in these areas such as thecustoms, the airports and the military branches.