ZiHao Wu

Mr. ZiHao Wu


Research Focus & Metrics

Academic Career Length

2Years (2025-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
22 January 2025 in Agriculture
Verified
Journal Article
Open access
PDF file
YLYun Liang
Yun Liang
WJWeipeng Jiang
Weipeng Jiang
YLYunfan Liu
Yunfan Liu
ZWZiHao Wu
ZiHao Wu
RZRun Zheng
Run Zheng
The citrus picking-point localization is critical for automatic citrus harvesting. Due to the complex citrus growing environment and the limitations of devices, the efficient citrus picking-point localization method becomes a hot research topic. This study designs a novel and efficient workflow for citrus picking-point localization, named as CPPL. The CPPL is achieved based on two stages, namely the detection stage and the segmentation stage. For the detection stage, we define the KD-YOLOP to accurately detect citrus fruits to quickly localize the initial picking region. The KD-YOLOP is defined based on a knowledge distillation learning and a model pruning to reduce the computational cost while having a competitive accuracy. For the segmentation stage, we define the RG-YOLO-seg to efficiently segment the citrus branches to compute the picking points. The RG-YOLO-seg is proposed by introducing the RGNet to extract efficient features and using the GSNeck to fuse multi-scale features. Therefore, by using knowledge distillation, model pruning, and a lightweight model for branch segmentation, the proposed CPPL achieves accurate real-time localization of citrus picking points. We conduct extensive experiments to evaluate our method; many results show that the proposed CPPL outperforms the current methods and achieves adequate accuracy. It provides an efficient and robust novel method for real-time citrus harvesting in practical agricultural applications.

5 February 2026 in Plant Phenomics
Verified
Journal Article
Open access
YLYun Liang
Yun Liang
RZRun Zheng
Run Zheng
NWNifa Wang
Nifa Wang
ZWZiHao Wu
ZiHao Wu
DWDong Wang
Dong Wang
YZYanxia Zhang
Yanxia Zhang
Root phenotypic traits such as length and number are critical indicators of plant growth and productivity. However, accurate extraction of these traits remains challenging due to the slender morphology, dense overlap, and frequent occlusion within root systems. Traditional digital image processing methods suffer from low throughput and limited robustness, while most deep learning–based approaches rely on semantic segmentation, which fails to distinguish individual roots and therefore limits their applicability in instance-level phenotypic analysis.To address these limitations, we propose Dual-Guided Asymmetric MP-Former (DGA-MP-Former), a novel instance segmentation model tailored for root phenotyping, with rice roots as a representative case. Building upon the MP-Former framework, our model introduces two key components: the Guided-Enhancement Pixel Decoder (GEPD) and the Asymmetric Dual-Query Decoder (ADQD). The GEPD enhances multi-scale feature representations via Hybrid Convolution Aggregator, Semantic-Guided Fusion Module and Frequency-Guided Feature Enhancement Module, effectively capturing fine root structures and low-contrast regions. ADQD employs asymmetric interaction between semantic and instance queries to improve long-range dependency modeling and instance separation in occluded scenarios.Additionally, we present the Rice Root Segmentation Dataset (RRSD), comprising of 343 high-resolution images with instance-level annotations. Experimental results show that DGA-MP-Former achieves state-of-the-art performance on RRSD, with 57.2% AP0.5:0.95 and 87.4% AP0.5. Importantly, the accurate instance segmentation results enable reliable computation of instance-level geometric traits, such as root perimeter and area. To quantitatively assess phenotypic measurement accuracy, Relative Area Error (RAE) and Relative Perimeter Error (RPE) are further introduced, achieving 26.4% and 20.2%, respectively. These results demonstrate that the proposed method effectively bridges instance segmentation accuracy and phenotypic quantification reliability, supporting high-throughput and precise root phenotyping.