ABSTRACT
In order to solve the problem of poor detection effect caused by color, size and picture noise of aluminum-plastic blister capsules, a method for detecting packaging defects of aluminum-plastic blister capsules was proposed based on multi-layer perceptron model. First, the batch number area of the drug board was used as the template, and the normalized product correlation gray level was used to match the legal position of the drug board to be detected. Then, the capsule blister region of the drug board was divided by the improved gray value projection method, and the gray value of the drug blister region, capsule volume and boundary characteristics were extracted. The multi-layer perceptron model was trained and tested to realize the identification of packaging defects such as missing, double cap and concave cap of aluminum-plastic blister drugs. Experimental results show that the improved horizontial-vertical projection algorithm can segment the capsule bubble cap region with 100% accuracy, high robustness and good segmentation effect. The identification accuracy of missing capsules, double caps and concave is high, and the average detection time is 4.47ms/ sheet, which can meet the quality inspection task of aluminum blister medicine board packaging.
- Zhao Baoshui. Research on key technologies for drug packaging inspection based on machine vision [D]. Jinzhou: Liaoning University of Technology, 2021.Google Scholar
- Wang Jun, Ge Bin, Chen Yinan, Application of improved NCC algorithm in aluminum-plastic pharmaceutical panel packaging defect detection [J]. Packaging Engineering, 2020, 41(17): 196-201.Google Scholar
- Gong Xiaoyun. Application of computer vision in drug packaging inspection[D]. Harbin: Harbin University of Science and Technology, 2019.Google Scholar
- Du Chunhua. Research on capsule defect detection and identification system based on machine vision [D]. Qufu: Qufu Normal University, 2019.Google Scholar
- WANG N H, TIAN S Q, ZHAO P, TRIZ Theory is Used to Improve Based on Gray Projection Algorithm for Electronic Image Stabilization[J]. Applied Mechanics and Materials,2011(1287):58-60.Google Scholar
- Su Yali, Wu Jianxing, Hui Wei, A spiking neural network for visual color feature classification [J]. Journal of Xi'an Jiaotong University, 2019, 53(10): 115-121.Google Scholar
- LIN Q Q, ZHOU J Z, MA Q R. Detection of the Fluff Fabric Surface Quality based on Machine Vision[J]. The Journal of the Textile Institute,2022,113(8):1668-1670.Google Scholar
- QI Q H, TIAN Y M, HAN L L. An improved image segmentation algorithm based on the maximum class variance method[J]. MATEC Web of Conferences,2020,309.Google Scholar
- Long S C, Huang X X, Chen Z Q, Automatic Detection of Hard Exudates in Color Retinal Images Using Dynamic Threshold and SVM Classification: Algorithm Development and Evaluation[J]. BioMed research international,2019(2019):6.Google Scholar
- Hou Jinyao, Gao Aihua, Liu Weiguo, Yan Lirong, Li Wenjin. Defect image segmentation algorithm based on morphological operations and image fusion [J]. Journal of Detection and Control, 2021, 43(05): 55-59.Google Scholar
- Xu Cong. Research on neural network recommendation algorithm based on graph structure learning [D]. Wuhan: Huazhong University of Science and Technology, 2021.Google Scholar
- Wang Yueqiao. Research on neural network training based on intelligent algorithms[D]. Guiyang: Guizhou University, 2021.Google Scholar
- Cheng Hong, Chen Wenjian, Sun Wenbang. An improved fast normalized product correlation image matching algorithm [J]. Optoelectronic Engineering, 2013, 40(01): 118-125.Google Scholar
- Li Gang. Research on liquor liquid impurity detection system based on machine vision [D]. Taiyuan: Taiyuan University of Technology, 2019.Google Scholar
- YU T, TIAN M Y. HOG features and connected region analysis based on workpiece object detection algorithm[J]. International Journal of Innovative Computing and Applications,2020,(11):2-3.Google Scholar
Recommendations
An efficient multilayer deep detection perceptron (MLDDP) methodology for detecting testicular anomalies with or without congenital heart disease (TACHD)
AbstractIn the real world, the detection of heart diseases is a challenging process. For detecting testicular anomalies with or without congenital heart disease, 2D ultrasound images through a computer-aided diagnostic support system are favorable. ...
Neural network based detection of hard exudates in retinal images
Diabetic retinopathy (DR) is an important cause of visual impairment in developed countries. Automatic recognition of DR lesions in fundus images can contribute to the diagnosis of the disease. The aim of this study is to automatically detect one of ...
Glaucoma detection using novel perceptron based convolutional multi-layer neural network classification
AbstractGlaucoma in retina is considered as a significant cause of irreparable vision loss. For the automated glaucoma detection on fundus images, several approaches have been recently developed. However, the extraction of optic cup (OC) boundary ...
Comments