Research on Test Conditions of Sealed-Relay Remainders Based on High-Speed Video Technology - Intelligent Computing for Sustainable Energy and Environment - page 507

test conditions are based on the movement of the remainders. Although this method
cannot accurately simulate the movement of remainders inside relays, the relationships
between test conditions and the output signals can be investigated and used as a basis
for improving the PIND test conditions. Therefore, the movement of remainders in
transparent rectangular cavities and relays with a transparent shell is researched in this
paper. In addition, there exist a wide range of remainders and test objects, therefore,
accurate estimation of motion parameters under all different conditions will not feasible,
and the research on test conditions should be based on statistics. In summary, the
proposed analysis method of remainder movements requires:

a ) Direct observation of the motion process of remainders in transparent cavity and
relays with a transparent shell, and the remainders are supposed to be larger than1mm
in diameter;
b ) Record of sufficient experimental images continuously and automatic
preprocessing is necessary;
c) Automatic calculation of the motion parameters of remainders and test objects.

Based on these, a movement-analysis system is developed based on high-speed video
technology. The system uses an existing testing platform of remainders and a sealed
transparent cavity. The system can pre-process the image of remainders with
appropriate image enhancement algorithms, calculate the location of remainders with
reference to the inner wall of the cavity; calculate the motion parameters of remainders
and test objects based on kinematic principle and plot the trajectory of parameter
changes. The overall diagram is shown in Figure 3. Here, the York phantom v7.3high-
speed camera with a storage space of 8GB is used in the analysis system.

2.2 Image Processing of Remainder Movements

1) Pre-processing and Filtering
According to preliminary experiments, the parameters of the camera were set as
follows: the resolution was 256 × 64, the pixel size was 22μm; the frame rate was
100000fps; the time of inter frame and exposure were set to 10μs and 8μs respectively
to ensure that the whole process of remainder motion can be captured. The ratio of pixel
size ρ is 46pixels/mm according to calibration.
In order to reduce the influence of color difference during image processing and
enhance image quality, the Gamma correction and grayscale processing and
enhancement are needed [12, 13]. The image pre-processing procedure of the system
is shown in Figure 4.
The noise will inevitably affect the image quality during the image acquisition
process, and image enhancement techniques are used for denoising. Three factors,
including calculation efficiency, filtering effect and edge effects, are important in
image enhancement. Therefore, six filtering methods were chosen in the research,
including median filter, Sobel operator, Prewitt operator, contrast enhancement filters,
adaptive filtering and Laplace Gaussian operator. The original grayscale images were
processed with the six methods and the results are compared as in Figure 5.From
Figure 5, we can find that the images processed with the median filter have the
clearest edge and less noise points. The median filtering algorithm denoises only after