Data Processing of Flir Tau2 Camera System in Microwave Insect Control Experiment

• Main Authors: Zongxiao Guo, 郭宗曉
• Other Authors: Kwo-Ray Chu
• Format: Others
• Language: zh-TW
• Published: 2017
• Online Access: http://ndltd.ncl.edu.tw/handle/x42b84

碩士 === 國立臺灣大學 === 物理學研究所 === 105 === Microwave insect control has been attracting academic attention for many years. There are many problems worth to research in the process of microwave insect control and one of them is data processing. This study is part of the microwave insect control project. In our microwave insect control, we set a 24GHz high power microwave source which is different from other low frequency microwave used before. In the 24GHz microwave insect control experiment, Flir Tau2 camera is set inside the chamber to detect the temperature field during heating. But during the temperature detecting , data output and data processing, we have to face a great number of problems. This study is mainly to solve two problems in data processing. Firstly, Flir Tau2 camera''s output images have noise which may cause by the complex electromagnetic environment or other unknown impact. Secondly, in order to process temperature data in batch quickly , it is necessary to accurately determine the object''s (grain) edge. In this study, two kind of image denoising algorithms (Median Filter and Gaussian Filter) are used to solve the first problem above. The advantages and disadvantages of this two kind of denoising algorithms are compared and we found that Median Filter can get a better denoising result. In order to solve the second problem above, we need to binarize the image at first step. We compare three kinds of binarization algorithms (Minimum Method, Intermodes Method, Otsu Method), and find that Minimum Method can achieve better results in this experiment. At the same time, a method is proposed to clearly identify the edge of object in some cases while the three method cannot work directly. After binarization, a batch data processing program is developed to complete the last step of data processing. Finally, a general data processing process is proposed in this experiment. Which can save considerable time and improve the accuracy of experiment.