Enhancing image quality in low-light conditions is a complex task due to challenges like poor contrast, blurred details, and noticeable noise. Conventional approaches often fall short in addressing these issues as they overlook the crucial blend of global and local insights necessary for effective image restoration. To this end, this paper introduces a low-light image enhancement method based on the synergy of global and local knowledge. Global information contextualizes overall brightness and background distribution, while local details enhance image clarity in specific regions. We innovatively introduce dilated depth-wise convolutions into the Transformer structure to achieve global information extraction. According to the local perception characteristics of the convolution structure, the convolution residual block is designed to extract local information. Then the dynamic focus fusion module is used to fuse the important part of global and local information. Experimental results on several commonly used benchmark datasets, both synthetic and real datasets, demonstrate that our proposed method achieves superior enhancement effects.