No abstract

Abstract

Abstract. It is good idea that machine vision is introduced to inspect the seasoning packets of instant noodles on-line. The image processing algorithm was a key when the vision system was built up to capture the image of dough with seasoning packets. Image of dough contour was analyzed to find if the dough has a perfect shape. Before pocketing instant noodle, several packets were checked if they lie on the dough by method of histogram statistics of brightness in HIS color model and threshold. Vegetable and oil packets were inspected by the histogram statistics of Hue while the silver salt packet was tested by the histogram statistics of luminance. Testing on actual production line, a system using the image algorithm has the precise identification rate of 100% and false detection rate of 99%, and the detection speed of better than 20 per minute. IntroductionInstant noodles represent a fast growing product and have a fast growing market in Asian countries. Since the competition becomes fiercer, instant noodles' profits is sharp decline and instant noodles enterprises confront with more severe challenge [1]. It is more important to improve the quality and varieties of instant noodles for meeting the requirement of different people. In automatic packaging production line, it is not a good way to arrange workers to control the quality of packaging instant noodle. As machine vision is introduced into industry field, the industrial camera is replacing the worker's eyes to observe and measure objects in the scene. Compared with the artificial examination, the machine vision has the merits as the high automaticity, sharp recognition capability, the accuracy measurement and has the widespread application prospect. With the development of machine vision technology and the lower cost of higher capability machine hardware, it is already becoming an important testing role in quality automatic detection about food and the agricultural product. Image processing algorithm is built up in time when images of scene in camera are complex and varied but stable in a special condition [2].In this paper, a visual inspection system was set up to be placed between the packaging position and packets slide position on the production line which consist of a holder with camera and light source to avoid the impact of existing production lines. The light source was a dome white light which can avoid nearly the noise light outside. The camera was a high speed color camera with 640x480 resolutions, 60 frames/s and 1394 interface. The system could capture the stable image though the dough moves fast down the camera. The algorithms of image process were tested in the system enough.

In view of the characteristics of our civil aviation, a fuzzy diagonal regression neural networks recurrent forecast model was proposed based on analyzing influential factors of passenger traffic volume. This model deals with the uncertain factors fuzzily and certainty factors using normalization in the front network layer, which solved the problem for inconsistent of importing dimension effectively. At the same time, Example proves the validity of the model. Practice proves that applying fuzzy diagonal regression neural networks recurrent forecast model to civil aviation passenger traffic volume is practicable, precise and universal, compared with the other methods such as the support vector regression, BP neural networks etc..

In view of the problems that the off-line inspection of traditional measuring instrument was inconvenient to automatically correct the Computer Numerical Control (CNC) machining process parameters and feedback control the in-process machining quality, the configuration collaborative monitoring method was proposed based on the analysis of the working principle and technological process of the CNC pneumatic measuring instrument. The real-time monitoring of the measuring instrument's measuring process was realized by means of the interaction between the configuration unit and the intermodule information. And then the pneumatic measuring instrument monitoring system was designed based on KINGVIEW and PLC, which can realize the real-time monitoring and controlling of measuring probe in manual mode, semi-automatic mode and automatic mode, overtravel and collision warning, procedure allowance feedback controlling, workpiece quality statistical analysis and so on. The system provides the basis for the monitoring of processing quality of batch manufacturing, as well as health maintenance equipment timely finding hidden trouble.