Research On Vision-based Inspection Technology For In-vitro Diagnostic Kit Production Line

This paper is closely based on the key technology of in-vitro diagnostic kitproduction line. A vision-based automatic inspection system is proposed for the needof quality detection. The structure and composition of inspection system is determined.The image processing and defect detection algorithms for different defects are studied.The following works have been accomplished.1. Considering the structure and assembly process of diagnostic kit, theinspection system is divided into3subsystems, which detect the strip defects, stripassembly defects and kit assembly defects respectively. The system arrangement inproduction line is confirmed.2. According to the detection requirements, the hardware and structural of thelighting system, image acquisition system and positioning system is designed. Thesoftware workflow and function modules are deduced.3. Size detection and location are based on the application of edge detection andstraight boundary fitting. Threshold segmentation and particle analysis were used todetect membrane surface defects. Experiment results showed that detection algorithmmet the inspection need of kit production line.4. Boundary detection and feature extraction are selected to detect stripassembly quality. Circle detection is used to judge kit assembly quality. Experimentresults showed that the inspection algorithm is a good substitute for human and theinspection accuracy fot the common assembly defects can reach98%.