The research team of the project for developing rose harvesting machinery for Kushui roses collected relevant data on-site

　　On May 25, reporters learned from the Lanzhou Academy of Agricultural Sciences that a flower picking machine jointly developed by teams from China Agricultural University and Tianjin Agricultural University has completed the visual recognition algorithm and the overall three-dimensional design. Prefabricated components have been used to assemble the first prototype, and the integrated data processing module is being debugged simultaneously. This marks phased progress in the research on mechanized picking of Kushui roses in Lanzhou.

　　In accordance with the arrangements and deployments of the CPC Lanzhou Municipal Committee and the Lanzhou Municipal People's Government, the Lanzhou Academy of Agricultural Sciences and relevant departments of Yongdeng County have coordinated and cooperated. Teams from China Agricultural University and Tianjin Agricultural University are accelerating the research and development of flower picking machines to break through the core technical bottlenecks in the picking of Kushui roses and boost the intelligent development of the rose industry. During this process, the professional team from China Agricultural University conducted on-site investigations at the Kushui rose planting bases in Yongdeng County. The Lanzhou Academy of Agricultural Sciences provided supporting data on local rose cultivation, information on the growing environment and analysis of the current industrial situation.

　　The team systematically studied the growth characteristics of Kushui rose, including plant architecture and flower distribution, and scientifically verified the technical feasibility of mechanized picking. It also comprehensively sorted out the operational difficulties, successfully developed visual recognition and spatial positioning algorithms specifically for Kushui rose, and completed the three-dimensional structural design of the fully automated picking machine. These efforts have established the underlying technical logic connecting visual perception and mechanical operation.

　　At present, at the Kushui Rose Planting Base in Yongdeng County, the team has carried out large-scale image data collection by assembling the first-generation prototype. The collection covers different lighting conditions, shooting angles and various flower forms. Meanwhile, standardized labeling rules have been formulated to classify the flowers, ensuring that the dataset can fully reflect the complex growing environment. The integrated data processing module is being debugged concurrently. (Report and photos by Zhang Wanhong)