Research On Machine-Vision-Based Online Detection Technology For Pharmaceutical Preparation

Medical injections are subject to various factors in theproduction process and can’t reach the quality standards for medical products.Medical injections are used widely on the clinical application. In order toensure the quality and safety of the medical injection, the medical injectionmust be tested carefully. Usually, the medical injection is tested by themanual inspection and the method has many disadvantages, such assubjectivity, low efficiency. The paper mainly focuses on the method ofMachine-Vision-Based online detection, and at last the method of onlinedetection is achieved.The method of Machine-Vision-Based online detection technology forPharmaceutical Preparation was designed. First of all, by analyzing theadvantage and disadvantage of online detection abroad currently, the contentof the paper combining with the china’s national conditions was made. Thenthe overall proposal was made and a hardware detection platform wasdesigned with many functions, such as communicating with the up positionmachine, clamping, rotating and stopping the bottle automatically, and then animage acquisition system which could collect the stopping bottle instantly andthen pass images to the up position system was designed. The average filterand region growing algorithm was used to handle the printer printing in caseto extract the corner of the model plane precisely to achieve the cameracalibration. At last, the software algorithm for the Machine-Vision-Basedonline detection was introduced. Details of the software based on dynamicthreshold were described.The host computer software based on the language of OpenCV and VC++was designed. To verify the accuracy of the method, the10ml pure solution ofsodium chloride and the solution provided by the factory which containsimpurities by the online method were checked. The result showed that the foreign matter was detected easily and quickly by using the device and thesoftware.