Search

Product appearance visual inspection is a research hot spot in the field of industrial machine vision technology. In the manufacturing process of Flexible Integrated Circuit Substrate (FICS), the defects of circle hole on FICS surface affect the electrical connection performance between chip and substrate, so detecting the defects of circle hole on FICS surface is an important content of visual inspection. However, the surface of FICS presents local high texture characteristics and the circle hole may have edge defects, which make it difficult to accurately detect the circle hole. To solve this problem, this paper proposes a circle detection method based on contour extraction and rule verification. Firstly, a contour extraction operator is used to extract candidate circular contours from FICS surface images. Secondly, the closed contours are divided into two categories according to the roundness of the contours, where the circle hole verification method based on the Helmholtz principle is used to verify large roundness contours, the iterative least squares method is used to verify small roundness contours. Finally, parameters of the resulted circular holes are calculated to determine whether they are defective circular holes or not. Compared with classical circle detection algorithms such as random circle detection and ED circle detection, the proposed method performs better results in terms of accuracy, precision and speed. The average accuracy and recall of the proposed method are 100%, and the average detection speed is 1.6ms on the test dataset.

An 8 weeks feeding trial was conducted to evaluate the effects of dietary n-3 LC-PUFA levels on growth performance, tissue fatty acid profiles and relative expression of genes involved in lipid metabolism of mud crab (Scylla paramamosain). Ten isonitrogenous diets were formulated to contain five n-3 LC-PUFA levels at 7 % and 12 % dietary lipid levels. Highest weight gain and specific growth rate were observed in crabs fed the diets with 19.8 and 13.2 mg g-1 n-3 LC-PUFA at 7 % and 12 % lipid, respectively. Moisture and lipid contents in hepatopancreas and muscle were significantly influenced by dietary n-3 LC-PUFA at the two lipid levels. The DHA, EPA, n-3 LC-PUFA contents and n-3/n-6 PUFA ratio in hepatopancreas and muscle significantly increased as dietary n-3 LC-PUFA levels increased at both lipid levels. The expression levels of Δ6 FAD and ACO in hepatopancreas increased significantly, and expression levels of FAS, CPTⅠ and HSL were down-regulated, with increased dietary n-3 LC-PUFA regardless of lipid level. Based on weight gain, the n-3 LC-PUFA requirements of S. paramamosain were estimated to be 20.1 mg g-1 and 12.7 mg g-1 of diet at 7 % and 12 % dietary lipid, respectively. Over all, dietary lipid level influenced lipid metabolism, and purified, high-lipid diets rich in palmitic acid reduced the n-3 LC-PUFA requirement of juvenile mud crab.

The surface defects of FPC are characterized by large scale variation and shape variation. How to quickly detect the surface defects of FPC has become a challenge in the quality control of the electronic manufacturing process. In this paper, we first develop an FPC surface defect method based on the Faster R-CNN object detection model. Based on the method, the FPN multi-scale feature fusion structure is then introduced, and the multiple receptive field fusion module (MRFM) is proposed to improve the ability of the model to extract large scale and multi-scale features. Finally, by embedding the multiple receptive field fusion module in the FPN multiscale output, we reduce the coupling between RPN and BBox head and improve the detection performance. To evaluate model performance, we build a dataset of FPC surface defects containing seven general defect types: short, open, pinhole, line damage, broken hole, exposed copper, and scratch. The experimental results show that the proposed model can achieve the mAP of 0.9557 and the mean recall of 0.9699 in the FPC surface defect dataset, which is better than the existing algorithmic model.

Product completion time prediction has important implications for make-to-order discrete manufacturing firms. In this paper, the product completion prediction problem coming from a practical company is considered, where the process data are with poor-quality, small number of features, category-dominated features, and low correlation between features and labels. Firstly, the characteristics of the provided process data is mined and preprocessed; and then a Deep Forest-based method is presented to predict the product completion time. Experimental results show that the method proposed in this paper achieves good performance.

Flexible printed circuit board (FPC) is a popular substrate for packaging integrated circuits (ICs). Detecting the circles rapidly on FPCs by using computer vision is very important to assess the quality of FPCs during its manufacturing. In this paper, a fast circle detection approach based on a threshold segmentation method and a validation check is proposed. In the algorithm, the image is firstly segmented by an adaptive heuristic threshold to obtain closed contours; then circle candidates are obtained by eliminating obvious non-circle contours; and at last, circle candidates are further validated to be circles according to Helmholtz principle. Experimental results show that the proposed method is very fast and has high detection rates to detect the circles on FPC images.

This paper presents an improved genetic algorithm based on niche strategy to solve the irregular nesting problem, which exists widely in modern manufacturing industry. For the niche strategy used in genetic algorithm, exclusion mechanism is introduced to avoid too many solutions with high similarity in the population, guaranteeing diversity of the solution set and preventing premature convergence of genetic search. No-fit polygon method based on Bottom-left strategy is used to evaluate the best placement position for every irregular part. Computational experiments based on data set published in the literature are taken to verify feasibility and effectiveness of the algorithm. The experiment results show that the proposed algorithm can be used to solve the nesting problem and generates some better solutions than the solutions published solutions for most examples.

For the wide application of CAD software in current electronic industry, the paper puts forward an idea of designing a kind of data interface between CAD and CAM systems, in order to achieve data sharing and data integration for various equipments in electronic assembly production. Taking component mounting/insertion machines as target equipments, this paper analyzes the data set to be processed by this data interface. Then we describe structure and format of data in a PCB source file generated by Protel, one commonly used CAD software, and discuss how to use the data interface to implement extraction and conversion of production data from a source file. The practical application proves that this data interface can greatly enhance the efficiency of the electronic assembly production.

In this paper, we discuss the process of low-volume and high-mix PCB assembly production in the electronic assembly industry and aim at an effective scheduling method to determine assembly sequence of different types of PCBs to reducing whole production time. The optimization problem can be modeled as an approximate Traveling Salesman Problem (TSP) when considering the effect of machine setup time on whole production time. An optimization algorithm based on Scatter Search method is proposed to search for good solutions to minimize machine setup time. The simulation results show that the proposed algorithm can output good scheduling solutions.