Automated parameter extraction for transmission towers relies on two main procedures: tower shape recognition and keypoint matching. Accurately determining structural parameters from an input tower image depends on identifying the tower’s archetype and aligning keypoints precisely. The effectiveness and precision of tower shape recognition and keypoint matching depend on the characteristics of the key-points’ feature descriptors. This paper introduces a new feature descriptor, based on Key Points Location Distribution (KPLD), which has a more compact structure. Initially, KPLD is used to match the point sets of extracted keypoints relevant to the transmission tower. Then, the tower shape identification process is based on observed differences in the matching outcomes. By adopting this new method, there is a significant improvement in the quality of keypoint matching for transmission towers. Moreover, this approach enhances both the efficiency and accuracy of tower shape recognition, providing a robust solution to the challenges in automated transmission tower parameter extraction.