Understanding cropland parcel change without producing cropland parcel maps

Cropland parcels are the basic units of agricultural production that define field morphology for a given area, and their size and shape can change substantially through land consolidation and other human activities. Monitoring these changes is important for understanding agricultural development and evaluating land management programs. Yet conventional approaches usually require complete cropland parcel maps to be produced at multiple time points and then compared, making the process time-consuming, computationally demanding and sensitive to mapping errors.

To that end, a team of researchers from China developed a new approach to identify whether and where cropland parcel changes have been undertaken without generating complete parcel maps. Their findings are published in the Journal of Integrative Agriculture.

“Instead of mapping every parcel, we compared the number of edge pixels within the same area at different times,” shares corresponding author Prof. Qiangyi Yu from Chinese Academy of Agricultural Sciences/Leibniz Institute of Agricultural Development in Transition Economies (IAMO), Germany. “When small, fragmented fields were consolidated into larger parcels, the total length of their boundaries generally decreases, leading to fewer edge pixels. The change in edge pixel counts can therefore serve as a proxy indicator for cropland parcel change.”

The researchers first extracted parcel edges from images and then removed false and redundant edges caused by factors such as buildings, agricultural textures and other sources of noise. They subsequently used the refined edge information to detect whether field morphology had changed.

“The approach was tested in five regions across East Asia—specifically in Central, North, East, and Southwest China, as well as Hokkaido, Japan—covering a range of regular (rectangular, square) and irregular (narrow, fragmented) field morphology,” says Yu.

The results showed strong and consistent performance. Across all test data, the method achieved an accuracy of 89.41%, precision of 89.23%, recall of 90.17% and an F1-score of 89.70%. All four measures exceeded 0.85 across the five study regions.

“The findings demonstrated that structural changes in cropland can be detected without producing complete parcel maps,” says Yu. “By focusing on changes in the number of edge pixels, this lightweight approach offers the potential to identify hotspot areas of cropland parcel structural change over large areas.”

"The concept could help identify hotspots of cropland change over large areas and support the monitoring of land consolidation programs,” adds co-corresponding author Prof. Baofeng Su from Northwest A&F University, China. “It could also be extended to regions experiencing cropland fragmentation, providing a new way to study how human activities reshape agricultural landscapes.”

Contact Authors:

Shiyao Li, E-mail: shiyaoli@nwafu.edu.cn; #Correspondence Qiangyi Yu, E-mail: yuqiangyi@caas.cn; Baofeng Su, E-mail: bfs@nwsuaf.edu.cn

Funder:

This work was jointly supported by the National Key Research and Development Program of China (2022YFB3903505) and the Central Public-interest Scientific Institution Basal Research Fund, China (Y2025YC87). The Land Systems Group at IAMO offered invaluable support throughout this research. This study contributes to the Global Land Programme.

Conflict of interest:

The authors declare that they have no conflict of interest.

See the article:

Li S Y, et al. 2026. Understanding cropland parcel change without producing cropland parcel maps: A novel structural change detection approach. Journal of Integrative Agriculture, 25(8): 3469–3482. https://doi.org/10.1016/j.jia.2025.10.014.