1 / 26

Two papers in IFAC14

Two papers in IFAC14. Guimei Zhang MESA (Mechatronics, Embedded Systems and Automation) LAB School of Engineering, University of California, Merced E : guimei.zh@163.com Phone:209-658-4838 Lab : CAS Eng 820 ( T : 228-4398). Sep 08, 2014. Monday 4:00-6:00 PM

Download Presentation

Two papers in IFAC14

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Two papers in IFAC14 Guimei Zhang MESA (Mechatronics, Embedded Systems and Automation)LAB School of Engineering, University of California, Merced E: guimei.zh@163.comPhone:209-658-4838 Lab: CAS Eng 820 (T: 228-4398) Sep 08, 2014. Monday 4:00-6:00 PM Applied Fractional Calculus Workshop Series @ MESA Lab @ UCMerced

  2. The first paper Paper title: AFC Workshop Series @ MESALAB @ UCMerced

  3. Motivation • This paper describes a software application for traffic sign recognition (TSR). • 2. The main difficulty that TSR (Traffic sign recognition) • systems faces is the poor image quality due to • low resolution, bad weatherconditions • or inadequate illumination.

  4. Overview of the proposed method Four stages: 1. image preprocessing • adjust the image size • a contrast limited adaptive histogram equalization is performed to enhance the contrast of the image • Transform the color image to grayscale image. • edge detection (by the Laplacian of Gaussian (LOG) filter). 2. Image segmentation Secondly, the traffic signs detection, where the ROIs (region of intersts) are compared with each shape pattern.

  5. Overview of the proposed approach 3. Thirdly, a recognition stage using a cross- correlation algorithm, where each traffic sign, is classified according to the data-base of traffic signs.( feature: normalize signatures) 4. Finally, the previous stages can be managed and controlled by a graphical user interface (GUI), which has been designed for this purpose.

  6. Example Imput image Grayscale image AFC Workshop Series @ MESALAB @ UCMerced 09/08/2014

  7. Overview of the proposed approach Laplacian of Gaussian function Edge detection

  8. Contour, its centroid and the starting point Regions of interest. Normalized signature of the ROI

  9. Shape pattern Normalized signature Normalized signature Shape pattern

  10. Imput image Rk: Cross-correlation matrix coefficient

  11. GUI First interface second interface

  12. Conclusion • A new traffic sign recognition system has been presented in this paper. • The image processing techniques used in this software include a preprocessing stage, regions of interest detection, the recognition and classification traffic sign, GUI designed. • The performance of this application depends on the quality of the input image, in relation to its size, contrast and the way the signs appear in the image.

  13. Discuss • Problems: I think there are some problems in this paper: • The feature is not robust to project transform. • Edge detection can be perform after image segmentation, maybe the efficiency can be improved. • Should add some contrast experiments, such accuracy and efficiency contrast with the existed methods.

  14. The second paper

  15. Abstract

  16. Materials and Feature extraction Experiment Material (plant) Sunagoke moss mat was used in this study Water content was determined as: where: tmw is the total moss weight (g) and idw is initial dry weight (g) of Sunagoke moss. Dry weight of moss was obtained by drying process in the growth chamber until there is no decrement in the weight of moss.

  17. Features: 1. Colour Feature (CFs: 22) • Textural Feature (TFs: 190) • Colour Co-occurrence Matrix (CCM) • 3. Back-Propagation Neural Network (BPNN) • A three layers BPNN performed better than the other type of ANN to describe the relationship between moisture content of the moss and the image features.

  18. Multi-Objective Optimization (MOO) 5. Neural Discrete Hungry Roach Infestation Optimization (N-DHRIO) algorithm

  19. The result of precision lighting system

  20. Conclusion • The intelligent machine vision for precision irrigation system using optimized feature selection has been developed. There is an improvement in optimizing feature selection using NDHRIO compare to the previous study. • The intelligent machine vision for precision LED lighting system has also been developed, and it shows effective to select LED light intensity which is appropriate to the certain part of the plant so that all parts of the plant can get enough light and proper intensity. • In large scale plant factory, those systems can optimize the plant growth and reduce the water consumption and energy costs.

  21. Discuss • In my opinion, if possible, we can improve it as follow: Many feature are employed to describe the object, though the authors proposed NDHRIO to select feature, the efficiency is an important issue. So I think we can first to use PCA( Principal component analysis) to reduce the featuredimension and improve recognition efficiency.

  22. Thanks

More Related