In order to improve the detection speed for lines in digital images, an improved detection method was proposed by combining the time-frequency domain transform and the spatial domain transform. Firstly, the wavelet lifting was used to extract the low frequency profile information and to restrain high frequency noises. Then, the gradient of a image was computed to obtain a binary image. On the basis of the principles that a line can be determined by two points and a line in the image is mapped to a point in the Hough Transform, the non-zero pixels were mapped into the accumulator cells with great probability instead of all accumulator cells following the detection sequence from the local to the global. Finally, the accumulator cells were counted to determine the parameters of lines in the image. It costs 213 ms to detect lines in the image with pixels of 128×128 by using the method proposed in this paper,which increases the detection speed for lines in images effectively.