Abstract: Sheep body morphometrics have been increasingly recognized as critical indicators for evaluating the growth performance, linear conformation, and genetic improvement in the livestock industry. Traditional manual measurement methods heavily relied on contact tools. These conventional approaches were characterized by low efficiency and high labor intensity, which induced severe stress responses in animals. The high-throughput and high-precision data demands of modern intensive animal husbandry could not be satisfied. Consequently, non-contact measurement systems based on computer vision have emerged as a cutting-edge paradigm for smart farming. In this study, a systematic and comprehensive review was presented on the current research progress of the non-contact sheep body size measurement technologies. The technological trajectory was traced from the two-dimensional machine vision to the three-dimensional point cloud reconstruction, as well as the emerging multimodal fusion frameworks. The data acquisition methodologies were extensively analyzed across different agricultural scenarios. The application matching, advantages, and technical bottlenecks of the mobile portable devices, fixed-channel systems, and fixed-arch measurement platforms were evaluated. These acquisition methods were examined in the context of both vast grazing pastures and intensive housing environments. The unmanned aerial vehicles were analyzed for the real-time tracking and spatial parameter estimation in the open pastures. And the deep comparative analysis was conducted on the measurement accuracy and algorithmic robustness between the linear parameters and the circumferential parameters. In the two-dimensional visual extraction schemes, the relative error of the linear traits was typically controlled within a low range, while the estimation of the circumferential traits suffered from significant errors due to the inherent loss of depth information and perspective distortion. Three-dimensional point cloud approaches demonstrated superior spatial geometric representation to significantly reduce the linear measurement error through the direct Euclidean distance calculations. Even with the added spatial depth, the circumferential measurements still remained a formidable challenge. Although the point cloud slicing and curve fitting algorithms like the cubic B-spline fitting, improved the accuracy, the precision was still severely compromised. The degradation was attributed to the self-occlusion of the sheep abdomen and inner thighs, as well as the non-linear expansion caused by the thick wool. Moreover, the evolution of the target segmentation and key-point localization was summarized from the conventional handcrafted geometric features to the data-driven deep learning models. The latest breakthroughs in the multimodal fusion technologies were also highlighted. Specifically, the synergistic application of the YOLOv12 instance segmentation and point cloud geometric fitting was examined for effectively decoupling the interference of the trunk distortion and complex backgrounds. Another application was given on the partly pose normalization mechanism, which was utilized to align the irregular postures and eliminate the nonlinear measurement errors. Despite these advancements, several critical bottlenecks were identified in the current measurement techniques. The dynamic adaptability of the algorithms was significantly degraded under continuous motion scenarios. There were severe motion blur and point cloud tearing in the fast-paced sorting channels. The algorithmic generalization was also hindered by the breed variations. The thick-wool breeds, such as the Tibetan sheep, suffered from severe key-point drift compared to the short-hair breeds, resulting in the failure of the geometric feature extraction. Finally, the future trends of the automated sheep body measurement are predicted to break through the existing technical barriers and the sensing paradigm is shifting from the single-source perception to the multi-modal data fusion. The integration of the depth cameras with the thermal imaging or solid-state LiDAR can adapt to the complex illumination and harsh farm environments. The deployment of the lightweight neural networks and edge computing architectures is urgently required to resolve the real-time processing bottlenecks of the massive point clouds. The instant pose normalization and parameter calculation can be enabled at the edge end. Moreover, the construction of the large-scale, cross-breed, and full-lifecycle open-source phenotypic databases is essential to provide the standardized benchmarks. The recent three-dimensional point cloud dataset for the Jining Qing goats can serve as a strong foundation to enhance the generalization capability of the fundamental models. The engineering integration of these measurement modules into the daily workflows of the sheep farms also represents the ultimate pathway. Highly protected sensors can be embedded into the smart feeding stations or automatic weighing-sorting gates to achieve the imperceptible, stress-free, and high-throughput morphometric monitoring. Overall, this review can provide a strong reference for the in-depth research and application of relevant technologies within the smart animal husbandry sector.