Received: 30-08-2022 / Accepted: 18-04-2023
This study aimed to assess the fruit trait diversity of cherry tomato germplasm through image-based phenotyping for the nutritious cherry tomato breeding program in Vietnam. Fruit images of sixty-eight cherry tomato accessions grown in winter crop season 2021 at the Vietnam National University of Agriculture were collected at the harvest stage and then were analyzed by SmartGrain software. Comparing the accuracy with the manual measurement using a simple linear regression model, the obtained results show that the fruit length and width extracted from the image had a high determination coefficient (0.81 and 0.76, respectively) and low root mean square error (2.95 and 2.35, respectively) compared with manual measurements. The image-extracted fruit traits were more closely and positively correlated with fruit weight as compared to manual measurements. The tomato germplasm were classified into 6 groups at the similarity coefficient of 65.95 from 5 image-extracted fruit traits. Fruit image-based phenotyping can replace the manual measurement and apply in tomato breeding, testing system, and further apply to other cultivars.