In order to improve the accuracy and efficiency of defect detection in power equipments， a time-division infrared polarization imaging system is constructed， combining a long-wave infrared detector and a polarizer. For electric transmission lines， multi-channel image acquisition is realized by controlling the polarizer rotation. Then， the polarization images with different channel numbers and different polarization angle intervals are evaluated using three parameters， i.e. root mean square error （RMSE）， peak signal-to-noise ratio （PSNR） and mean structural similarity （MSSIM）， respectively. The experimental results show that as the number of channels increases from 3 to 18， the image quality is gradually improved， especially in the range of 4 to 9 channels， where the improvement of imaging quality is most significant. However， further increasing the number of channels does not have a significant effect on the image quality， which leads to redundant calculations； for 3-channel polarization images， small angular intervals are able to capture the detailed information of fine targets， while large angular intervals are able to effectively separate the target from the background and reduce the noise interference， which are both suitable for some specific application scenarios. The experimental results are instructive for the polarization channel selection strategy of infrared polarization imaging.