Utilizing the rapid switching capability of imaging devices such as projectors, binary patterns are widely employed in fast 3-D imaging. However, existing binarization methods typically require extensive peeling to eliminate encoding noise and achieve approximate sinusoidal patterns. This behavior inevitably results in the use of 3-D measurement systems operating at shallower depths. In 3-D surface measurement, the accuracy of phase extraction from captured fringe images can be compromised due to low signal-to-noise ratios. To address this, our paper introduces a weighted binary encoding approach that leverages intensity information for practical 3-D reconstruction. Our approach involves the allocation of information discarded by the original temporal-spatial binary (TSB) encoding method, assigning each code in the binary sequence to an intensity interval. A sinusoidal fringe pattern, encoded in an 8-bit format, is segmented into N single-bit binary patterns. These patterns are then projected onto the target object under conditions of nearly in-focus, thereby creating approximate sinusoidal fringes. Detailed analysis of the encoding process in our paper demonstrates our higher utilization of intensity information. Through comparative experiments encompassing planar measurements and composite object measurements, we demonstrate that this newly proposed encoding strategy offers superior phase and 3-D measurement accuracy compared to existing methods while also featuring scalability.