Against the backdrop of precision agriculture and the development of intelligent agricultural machinery, current domestic monitoring systems for wheat combine harvesters are plagued by limited functionality, low intelligence, significant errors in parameter monitoring, and yield estimation results prone to inaccuracies. Specifically, they lag behind mature international systems in terms of fault warning accuracy, data transmission efficiency, and yield visualization capabilities. This study seeks to realize comprehensive and precise monitoring, reliable fault early warning, and intelligent yield prediction for wheat combine harvesters across all operating conditions. To this end, it innovatively adopts CAN bus integration technology and impulse-type grain flow sensors to develop a comprehensive system for monitoring the operational status and warning faults of wheat combine harvesters, which covers the entire operational process. By integrating GPS positioning, multi-sensor parameter acquisition, and intelligent analysis modules through CAN bus integration, the system enables unified monitoring of geographic information, operational data, cleaning loss, and fault status. Additionally, it incorporates a yield measurement module based on an impulse-type grain flow sensor to generate the real-time yield distribution maps. Field experiments demonstrate that the system achieves an alarm accuracy of 97.3%, controls the fuel consumption measurement error within 5%, and limits the relative error of yield measurement accuracy to no more than 4%. Notably, the impulse-type grain flow sensor exhibits stable static detection accuracy and rapid, precise dynamic measurement performance—laying a solid foundation for the automation and intelligent advancement of combine harvester technologies.

Lodged crops have posed the significant difficulties on in the traditional harvesting. The crops can also fall over after heavy rain or strong winds. This research aims to enhance the operational precision and efficiency of unmanned combine harvesters, particularly in the challenging task of the lodged crops. A reliable Light Detection and Ranging (LiDAR) system was developed to real-time monitor and accurately detect the plant lodging in rice and wheat fields. An acquisition system of the crop height was also developed to efficiently harvest using LiDAR and tilt sensors. Firstly, the high-resolution 3D point cloud data was captured from the LiDAR sensor, and then transformed into a ground coordinate system using rigid transformation algorithms. An accurate spatial representation of the crops was obtained relative to the harvesting machine. Secondly, the region of interest (ROI) was defined, according to the LiDAR's scanning angle and the cutting width of the combine harvester. The data processing was immediately performed on the area ahead of the harvester, in order to optimize the computational efficiency. Noise filtering was then used to eliminate the outliers for the high data quality. The point cloud data within the ROI undergoes was then optimized after noise reduction. A grid grid-down sampling on the XOY plane was carried out to convert the 3D point cloud into a 2D rasterized model of the crop canopy. Subsequently, a canopy height model (CHM) was generated using Delaunay triangulation. A triangulated surface was formed to accurately represent the height variations of the crop canopy. Finally, the real-time adaptability was maintained, as the harvester moved in the field. A dynamic meshing approach was implemented to integrate the unprocessed points into the existing Delaunay triangulation network. The CHM was continuously and accurately updated in real real-time. The high precision was maintained during dynamic updating, as the machine encountered the varying lodging in the field. Such a high level of accuracy was suitable for the real fields. Furthermore, the intersection-over-union (IoU) metric was averaged 95.26% in the identification of the lodging areas. As such, the excellent performance was achieved to accurately delineate the boundaries of lodged crop regions. This metric was fully met the requirements of the high precision of the detection on of the target areas during harvesting. Additionally, each frame of the point cloud data was processed in an average of 0.875 s, indicating the real-time monitoring and rapid adjustments to the harvesting path and operations. The operational efficiency of the processing speed were was obtained in the dynamic scenario. In conclusion, the real-time monitoring was realized to significantly improve the precision and efficiency of unmanned combine harvesters in the lodged crops. Advanced spatial transformation, noise filtering, grid down sampling, and dynamic Delaunay triangulation were integrated to for the high accuracy and robust performance under varying field conditions. The field experiments also validated the practical applicability using LiDAR. The finding can offer a reliable solution to reduce the operational losses for the higher crop yields in modern agriculture. Overall, the more efficient and accurate harvesting can be realized to enhance the precision farming using unmanned harvesting machinery, particularly for the lodged crops in many fields.