To suppress the high-level noise of raw images from the low-light image sensor, this paper proposes a collaborative filtering algorithm based on exact noise variance of transform domain. Firstly, the noise of low-light-level images are modeled as Poisson-Gaussian mixed noise and are performed by variance stabilizing transformation. Secondly, a calculation method of exact noise variance is proposed in 3-D transform domain based on L1-TGV regularization. Finally, the denoised images are obtained by embedding the exact noise variance into BM3D algorithm to improve patch matching and shrinkage accuracy. The proposed method and the other four noise suppression methods are used to process the same unnaturally degraded images and the raw low-light-level images, respectively. Numerical experiments on unnaturally degraded images express that the proposed method can effectively remove high-level noise and maintain image textures. Compared with BM3D algorithm, the proposed method can improve the peak signal-to-noise ratio by up to 2.15dB and the structural similarity by up to 0.106, respectively. Moreover, the testing of the raw low-light images further confirms the best performance of visual effect in contrast with the other four methods.