Digital feature selection. The selection of digital
feature depends on the trait of the display on SLM. SLMs normally use seven segments display which have seven strokes for each digital. The different combinations of these strokes can express different digitals. Through the comparison with digital template, the digital can be recognized. Therefore the strokes can be used as the feature of a digital. In practice, the strokes of each digital template are encoded into binary code so that only the binary codes need to be compared. Strokes confirmation. Strokes of each digital are
confirmed by sampling on the binarized image. When all the strokes of each digital are confirmed, the binary feature code of the digital is obtained. During the sample stage, if the sampled white pixels are more than the preset threshold in the stroke area, then it is judged that the stroke exists.
EXPERIMENTS
To verify the validity and reliability of the automatic calibration system, we did a lot of comparison experiments. The comparison is between manual calibration and automatic calibration. We calibrated the same SLM in these two ways with the same times. The experiment results show that the measured average value and type A measurement uncertainty are almost the same, but the calibration time in automatic way is shortened by 50%.
In addition, we test the performance of the image recognition method. The result is shown in Table II. Images of complicated background can be recognized with satisfied result. If the light intensity and incident angle are adjusted properly, the correct recognition rate will be improved.