The problems of low light image noise and chromatic aberration is a challenging problem for tasks such as object detection, semantic segmentation, instance segmentation, etc. In this paper, we propose the algorithm for low illumination enhancement. KinD-LCE uses the light curve estimation module in the network structure to enhance the illumination map in the Retinex decomposed image, which improves the image brightness; we proposed the illumination map and reflection map fusion module to restore the restored image details and reduce the detail loss. Finally, we included a total variation loss function to eliminate noise. Our method uses the GladNet dataset as the training set, and the LOL dataset as the test set and is validated using ExDark as the dataset for downstream tasks. Extensive Experiments on the benchmarks demonstrate the advantages of our method and are close to the state-of-the-art results, which achieve a PSNR of 19.7216 and SSIM of 0.8213 in terms of metrics.