| China is a large liquor and beverage producer, among these liquid products, liquid products filled in irregular-shaped bottles occupy considerable proportion. Greater attention has been given to its quality security problem. Now, the manufacturers have supplied a large amount of manpower and materials, the traditional manual testing cannot meet the need of high speed and quality in mass production. Chinese Jing wine contained in irregular-shaped bottle is taken as the subject in this paper, the relevant image inspection technology is expatiated to fulfill the task of online and automatic test of the quality of health wine. The successful development of the equipment proposed in this paper will greatly enhance the production efficiency and automatic level of manufacturers. It will have the enormous social meaning and the economic efficiencyFirstly, this paper elaborated the background information of this paper’s research, current situation of manual testing and the research and application of visible particles test in liquid done by related institutions both at home and abroad. Then the definition of machine vision and the compositions of vision system are introduced, meanwhile the application of machine vision in all aspects of our life and production is also presented.After elaborating several key technologies of testing system based on machine vision, the detailed plan of testing system is designed combining with the analysis of difficulties of the tested object in this paper. Mechanical system and the bottle-filled line should be connected perfectly. When the product enters into the testing system through production line, the irregular-shaped bottle will be turned over-stopped by the mechanical hand according to scheduled plan. The particles in the bottle continue to move under the influence of inertia and gravity. A series of pictures will be taken by cameras mounted on swing arm. Then the particles can be identified by the analytic result of those pictures. In the end, the inferior products with particles will be removed in the classification field. This plan only uses one pendulum arm to complete the black, the white particles tracking inspection, easy to control, and reduced the cost greatly.Secondly, the procedure of image processing of particle testing is presented while the algorithm of visible particle testing is discussed. The positioning technique in image preprocessing is also presented. After the moving object is tested by improved difference algorithm to find the doubtful point, different processing methods are applied according to the size and shape of the particle. The particle with larger size and slender shape can be identified by Blob analogy. The particles that cannot be identified by size, shape from bubble will be treated with tracking algorithm proposed in this paper, particles can be identified by the difference of directions of motion. The experimental result indicated that the particles inspecting algorithm in this article surpass manual testing.At last, the overall software of testing system of visible particles in irregular-shaped bottles is developed. Each functional module is performed after key techniques such as image acquisition and processing, reliable communication are solved. |
PDF Full Text Request