Plasma image edge detection based on the visible camera in the EAST device
© The Author(s) 2016
Received: 28 April 2016
Accepted: 18 November 2016
Published: 1 December 2016
The controlling of plasma shape and position are essential to the success of Tokamak discharge. A real-time image acquisition system was designed to obtain plasma radiation image during the discharge processes in the Experimental Advanced Superconducting Tokamak (EAST) device. The hardware structure and software design of this visible camera system are introduced in detail. According to the general structure of EAST and the layout of the observation window, spatial location of the discharging plasma in the image was measured. An improved Sobel edge detection algorithm using iterative threshold was proposed to detect plasma boundary. EAST discharge results show that the proposed method acquired plasma position and boundary with high accuracy, which is of great significance for better plasma control.
KeywordsEAST Tokamak Plasma image Sobel algorithm Edge detection
The calculation of EFIT (including offline EFIT and RTEFIT) requires a large number of electromagnetic measurement data from magnetic probes, flux loops, etc. The plasma-forming process is quite complex in the Tokamak start-up phase. The start-up plasma is not in equilibrium state, and the result of magnetic measurement hardly reflects the real situation of plasma current because of the existence of induced eddy current in the vacuum chamber wall (Liu et al. 2008; Zhou et al. 2015). Thus, the calculated result from EFIT is inaccurate and unusable for reliable plasma current and its position control. In addition, during the discharge process, the distance from the outer closed magnetic surface to the inner wall of mid-plane of the high field side (also called as Gap) is a very important parameter. During the preliminary discharge stage, the Gap influences the initiation of plasma, and accurate Gap control becomes helpful for reliable plasma current and position control. EFIT is also one of the methods to get Gap (Qian et al. 2009). Because the Gap will influence the initiation of plasma, it is more meaningful to obtain the accurate distance between plasma and inner wall of the EAST.
A fast visible camera, as a kind of imaging device, can obtain the radiation images in real-time during the plasma discharge (Jia et al. 2015; Yuan et al. 2013; Chapman et al. 2014). This paper mainly discussed how to acquire the plasma image and detect the spatial position of plasma boundary in EAST. A fast image acquisition system for EAST was introduced in this paper. A Sobel edge detection method with improved iterative thresholding algorithm was proposed in real-time plasma boundary detection. According to the EAST device structure, spatial location in the image was calibrated. Gap can also be obtained. It has considerably practical meaning for further plasma control.
The rest of the present paper is organized as follows. According to the structure of the EAST device and the position of the observation window, a fast camera image acquisition system and the camera calibration were presented in second section. In third section, the plasma image edge detection algorithm is described in detail. The experimental results and discussions are given in fourth section. Finally, the conclusion is drawn in fifth section.
Structure of the EAST visible camera system
The visible camera set-up
The visible/IR endoscope system including the visible camera, the IR camera and the spectroscope has been installed on the EAST device for the first time in 2014 EAST campaign. The IR camera used in this system is a FLIR SC700BB (2.5–5.0 μm IR ranges). The maximum frame rate is up to 2.9 kHz with a 132 × 3 pixels sub-window and 115 Hz for full-frame (640 × 512 pixels). The visible camera is a Phantom V710, and the maximum frame rate at full resolution (1280 × 800 pixels) is up to 7530 Hz. The spatial resolution of this system along the divertor plate in the poloidal direction is 4 mm for the IR camera and 3 mm for the visible camera. The visible/IR integrated endoscope system can monitor discharge process and the temperature distribution of the first wall and the divertor targets in real-time. In this paper, the authors focus on the observation of the visible camera system.
Hardware construction of the visible camera system
The CCM is the central control machine in the operation of the EAST experimental device, which is mainly responsible for the unified control and management of all subsystems in the experiment. During the discharge process, the CCM supplies shot number and acquisition time to image acquisition machine and sends trigger signal to the visible camera. The image acquisition machine is responsible for image acquisition, image processing and data uploading to the data server machine.
When an experiment shot starts, the visible camera begins to wait for a trigger from the CCM. Once the trigger arrives, the visible camera begins to capture images and get information about the discharging plasma by real-time image processing. When the shot ends, the camera system immediately sends images and the processing results to the data server machine for further analysis by physical researchers. The data server machine has a huge storage device for storing discharging images and processing the results.
Software design of the visible camera system
V710 camera initialization. Parameters such as sampling rate, exposure time, contrast ratio, are set to proper values.
UDP monitoring. UDP monitoring process is created to receive UDP data from CCM, including shot number and acquisition time.
Waiting for the trigger from CCM. The V710 camera waits for the start signal before image acquisition. Once the trigger signal arrives, the system starts image acquisition.
Video storage and image processing. Discharge video is saved and real-time processing is performed to detect plasma image edge, compress and save images.
Uploading image data. When a discharge is finished, the system uploads the image data to the data server machine. If the discharge continues, the system returns to step (2) and repeats UDP monitoring for the next discharge. Otherwise, the system exits.
The visible camera calibration
Due to the influence of image quality, the maximum pixel number in a grid is 96, and the minimum is 92. So d max = 2.92 mm, and d min = 2.86 mm.
The maximum pixel deviation of the actual distance is 0.06 mm. By calculation, the maximum number of pixels between the outer closed magnetic field surface and the inner wall of the high field side of the device is about 40–50, so the maximum error is about 2.40–3.00 mm. Based on all statistical pixel numbers, the average size of a pixel is 2.88 mm.
Plasma edge detection
As for image edge detection, it is very important to choose a suitable threshold value (Wang et al. 2005; Obrien et al. 1993). This paper applies a modified iterative algorithm to obtain the most reasonable threshold value, which can divide an image into two parts as background and goal. Then a modified Sobel algorithm is applied to complete the plasma edge detection.
The image threshold segmentation
Before processing threshold segmentation, some pre-processing steps for the original image are required. First, the original image is changed into gray-scale image. Then, noise points made by stray light and camera noise in the image are removed by weighted median filtering method.
If the histogram of the image has two obvious peaks, we can quickly get the satisfactory result. But the original images probably have various disturbances, such as shadow, uneven luminance, and different contrast ratio. If a fixed threshold is selected for image segmentation, the result may not be acceptable for all images. So we proposed an improved threshold algorithm, which can select different threshold based on image features during plasma discharge.
Calculate the maximum gray value P max and minimum gray value P min, and set the initial threshold value t 0 = (P max + P min)/2. According to t 0, segment the general position during the plasma discharge and regard it as the target image.
According to threshold value t 0, segment the target image into foreground and background, and then calculate the average gray value E 0 and E b .
Calculate the new threshold value t 1 = (E 0 + E b )/2.
If t 1 = t 0, the calculated value is the best threshold value. Otherwise, go back to step (2), change t 0 to t 1 and continue the iteration until t k+1 = t k .
Use the best threshold value t k or t k+1to make the image binarization and then output the processed image.
Edge detection using the improved Sobel algorithm
After threshold segmentation of the image, the improved Sobel algorithm is used to detection the plasma boundary. The principle of Sobel algorithm is as follows:
Experimental results and discussion
The visible/IR cameras system has been put into use in EAST campaign. With the above-mentioned algorithm, the EAST discharge plasma image is identified and it can get excellent results. It mainly identifies circular plasma during the start-up state and the high field outer closed magnetic surface under configuration in divertor. Combining with the calibration of space position of the image, the time evolution of the distance (Gap) is also obtained. It can provide reference to plasma shape control. The improved plasma position control system has been designed and implemented using the fast CCD. The future control system should solve the Tokamak circuit equation using appropriate models to establish the horizontal and vertical magnetic fields, which will be sequentially compared to the optimal values. The control algorithm should also be studied to solve the current in the poloidal field coils by adjusting the magnitude and changing rate of the current of poloidal coils to ensure the steady-state operation of Tokamak plasma. In addition to the present hardware and software setups, a real-time computer is also needed for the acquisition of electromagnetic measurement data, and the control instructions should be sent to the poloidal field coils power system through real-time computing and data processing, fast enough to meet the demands of plasma control.
The visible camera image acquisition and processing system is a practical system for real-time monitoring and for the controlling of the plasma discharge process, which is essential to the tokamak experiments. According to the structure of the EAST device and the position of the observation window, the fast camera image acquisition system was designed. The visible camera and infrared camera share the same window to observe the plasma discharge. The visible camera system is described in detail. As for the plasma edge detection, first pre-processing is performed to eliminate noise. Then the modified iteration threshold value is used to eliminate background and other useless information. After that, the improved Sobel algorithm is used to detect the configuration of plasma, and the plasma boundary location is obtained through the visible camera calibration, which can provide reliable data for feedback control during long-time plasma discharging. The experiment results during the EAST discharge show that the method of acquiring plasma position and boundary proposed in this paper has a high accuracy, which plays great importance on the EAST plasma position control.
SS designed the structure of the visible camera system, analyzed the edge detection algorithm, and wrote the manuscript; CX designed the hardware and the software of this visible camera system; MC and ZY calibrated the visible camera system in the EAST, and had done experiments, and analyzed the experimental data. All authors read and approved the final manuscript.
The authors are grateful to all members of the EAST team for their contribution to the experiments. This work is supported by the National Natural Science Foundation of China (Grant No. 11105028), and the National Magnetic Confinement Fusion Science Program of China (Grant Nos. 2013GB102001, 2015GB102004).
The authors declare that they have no competing interests.
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.
- Chapman IT, Holgate JT, Ayed NB et al (2014) The effect of the plasma position control system on the three-dimensional distortion of the plasma boundary when magnetic perturbations are applied in MAST. Plasma Phys Controll Fusion 56(7):075004View ArticleGoogle Scholar
- Jia MN, Yang QQ, Zhong FC et al (2015) A tangentially visible fast imaging system on EAST. Plasma Sci Technol 12:991–995View ArticleGoogle Scholar
- Liu CY, Xiao BJ, Wu B et al (2008) Modeling of first discharge in EAST tokamak. Plasma Sci Technol 10:8–12View ArticleGoogle Scholar
- Obrien DP, Ellis JJ, Lingertat J et al (1993) Local expansion method for fast plasma boundary identification in JET. Nucl Fusion 33:467–474View ArticleGoogle Scholar
- Qian JP, Wan BN, Lao LL et al (2009) Equilibrium reconstruction in EAST tokamak. Plasma Sci Technol 11:142–145View ArticleGoogle Scholar
- Qian JP, Weng PD, Luo JR (2010) Technical diagnosis system for EAST tokamak. Fusion Eng Des 85:828–835View ArticleGoogle Scholar
- Wan BN (2009) Recent experiments in the EAST and HT-7 superconducting tokamaks. Nucl Fusion 49(10):104011View ArticleGoogle Scholar
- Wan BN, Li JG, Guo HY et al (2013) Progress of long pulse and H-mode experiments in EAST. Nucl Fusion 53(10):104006View ArticleGoogle Scholar
- Wang ZT, Mao GP, Yang QW et al (2005) Identification of plasma boundary and position for HL-2A tokamak. Plasma Sci Technol 7:2905–2907View ArticleGoogle Scholar
- Wu ST (2007) An overview of the EAST project. Fusion Eng Des 5:463–471View ArticleGoogle Scholar
- Xiao BJ, Yuan QP, Humphreys DA et al (2012) Recent plasma control progress on EAST. Fusion Eng Des 87:1887–1890View ArticleGoogle Scholar
- Yuan QP, Xiao BJ, Luo ZP et al (2013) Plasma current, position and shape feedback control on EAST. Nucl Fusion 53(4):043009View ArticleGoogle Scholar
- Zhou ZB, Yao DM, Cao L (2015) The upgrade of EAST divertor. J Fusion Energy 34:93–98View ArticleGoogle Scholar