Analysis and Elimination Method of the Non-Fixed Column Noise for Quantum Image Sensor

In order to eliminate the non-fixed column noise of quantum image sensor (QIS), a gray value compensation method based on guide filter is proposed in this paper. According to the analysis of generation principle, the noise is introduced by the inconsistency of 1-bit ADC transition points and the readout noise of pixels. First, the raw image is acquired by the sensor model. Then the high-frequency non-fixed column noise components are separated from the image details. At the same time, the mean value of the noise components is calculated to obtain the noise vector of the non-fixed column. Finally, the non-fixed column noise vector is subtracted from the raw gray value of the corresponding column for correction. After the column noise is eliminated, the performance evaluation shows that the SDCMV of 128 grayscale decreases from 2.6978LSB to 0.9016LSB, and the SSIM increases from 0.8028 to 0.8641. It is proved that the proposed method can effectively eliminate the non-fixed column noise of QIS, while maintaining image details.