The Prototype Research Of The DC Picture-taking System Based On Photogrammetry Technology

In order to overcome the shortcomings of the old method of picture-taking, such as the low speed of picture-taking, low efficiency, off-grade pictures and low percent of pass, so the DC picture-taking system has been developed. In DC picture-taking system which is the reform of the old one, the collecting of the facial graphs is easier, the transmitting of the pictures can be through the computer in time, controlling the system on line, and the pictures can be cut or scaled so as to meet the requirements of size and legibility.The DC picture-taking system based on photographic surveying Technology is designed for collecting the qualified picture of the face and adopts the design idea of Face recognition. Face recognition, a kind of biology identification, is a new subject developed these years which has significant scientific value on research as well as practical value. Because of the acceptable advantages of directness, convenience and non-infringement, the system has a great application in figure identification and mode identification field. Moreover, the system is the improvement and upgrading of the form DC picture-taking system. After a series of the image preprocessing, the system collect the characteristic parameters of a person's face, judge and classify these collected characteristics through using the covariance analytical skill of the artificial neural networks, then cut and scale the original pictures automatically according to some certain requirements. This system, utilizing the artificial neural networks technique, introducing fuzzy theoretical knowledge simultaneously, adding fuzzy logic into the traditional rate neural networks structure of hierarchical mode, adopting the fuzzy inference rule and structure in the neural networks, attempts to structure the net works and construct fuzzy neural networks. Lots of techniques related to graph processing, such as:gray level thresholding, binary image, image automatic adjustment, remove discrete white noise, scaling image, thinning, curve smoothing, etc. has been used in this system. In addition, the technique that the fuzzy neural networks analyze, judge, and classify from the collected characteristic parameters is also been applied in the system. The experiences prove that the system has a great practical value.