Abstract: Accurate phenotypic analysis is essential for the conservation, breeding, and optimization of Camellia Oleifera, a vital economic tree species widely cultivated for its oil-rich seeds. Precise measurement of phenotypic parameters plays a critical role in advancing breeding programs, improving crop yields, and optimizing genetic traits. However, the complex canopy structure of oil tea trees presents significant challenges for extracting phenotypic data under field conditions. Existing methods, such as traditional manual measurements and multi-view stereo (MVS) reconstruction techniques, often face limitations in accuracy, efficiency, and scalability, particularly when applied to dense and intricate tree canopies. To address these challenges, this study introduces an innovative approach combining neural radiance fields (NeRF) and path analysis for the high-precision extraction of phenotypic parameters from field-grown Camellia Oleifera trees. This approach utilizes multi-view images captured from fixed positions around the tree, and a NeRF model is trained to reconstruct an accurate three-dimensional point cloud of the tree, preserving both structural details and high-frequency characteristics. Path analysis is subsequently employed to segment the tree stem and leaf clusters within the point cloud, enabling precise extraction of key morphological and phenotypic parameters. Experimental results show that the NeRF method significantly outperforms traditional MVS methods in terms of both efficiency and reconstruction quality. Specifically, the NeRF approach reduces the point cloud reconstruction time by approximately 90%, while also improving reconstruction quality, with the peak signal-to-noise ratio (PSNR) increasing by 10%. Furthermore, the path analysis-based segmentation method demonstrates superior performance in separating stems from leaves, with improved recall, precision, F1 score, and overall segmentation accuracy compared to conventional geometric feature-based and region-growing methods. The segmentation time is reduced by about fourfold compared to the geometric feature method and by approximately threefold compared to the region-growing method. For phenotypic parameter extraction, the proposed method achieves high accuracy with minimal error compared to manual measurements. Specifically, the errors in tree height, crown width, canopy height, and trunk diameter are 0.519%, 0.325%, 0.364%, and 4.491%, respectively. Regarding leaf parameters, the coefficients of determination (R²) for leaf length, leaf width, leaf area, and leaf shape index are 0.98, 0.94, 0.97, and 0.93, respectively, indicating excellent agreement with manual measurements. These results demonstrate that the proposed method is capable of accurately reconstructing three-dimensional point clouds of Camellia Oleifera and extracting precise phenotypic parameters even under challenging field conditions where complex canopy structures are involved. This approach provides a promising solution to the long-standing challenges of phenotyping in plants with dense and complex canopies and offers a fast, reliable, and cost-effective alternative to traditional measurement methods. The ability to quickly and accurately obtain detailed phenotypic data for large numbers of trees in field settings opens up new possibilities for improving breeding strategies, supporting precision agriculture, and advancing plant phenotyping research. Additionally, this method could have broader implications for other species with similarly complex canopy structures, potentially benefiting forestry, horticulture, and other sectors. Future research will focus on refining the algorithm to handle a broader range of plant species, improving computational efficiency, and exploring the integration of this methodology into real-world agricultural applications for enhancing crop management, yield prediction, and precision farming.