Vision-Based Agricultural Operation Mapping for Mixed Machinery
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current agricultural operation monitoring systems are inadequate for coordinating multiple machines, especially older machinery and towed implements without GNSS or communication capabilities, and are unreliable in varying operating conditions.
Innovation Solution
A system utilizing imaging sensors mounted on agricultural machines to obtain image data, analyze it with a detection model, and classify objects, identifying working machines to log and update operational information in an operational map, even for machines without positioning or communication systems, ensuring redundancy and validation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a fully connected system with GNSS and communication links is used to monitor and track agricultural machines, then monitoring capability and coordination are improved, but system complexity increases and reliability decreases due to communication failures and lack of support for older machinery
Solution Approach 1:
The patent introduces imaging sensors as an intermediary mechanism to capture visual data of agricultural machines and their operations. This mediator enables monitoring without requiring direct communication links between all machines, thus reducing system complexity while maintaining reliability. The imaging sensors capture images that are then processed to extract operational information, providing a robust monitoring solution that works independently of communication network stability.
Solution Approach 2:
The patent replaces the mechanical/communication-based monitoring system (requiring GNSS receivers and data networks) with an optical system using imaging sensors. This substitution eliminates the need for complex communication links and GNSS hardware on older machinery, as the imaging sensors capture visual information that can be processed to determine machine positions and operations, thereby reducing system complexity while improving compatibility with diverse equipment.
2Measurement precision
If GNSS capability is installed on all agricultural machines for tracking, then positioning accuracy is improved, but cost increases and adaptability decreases due to inability to support older machinery
Solution Approach 1:
The patent makes the imaging sensor system universal by enabling it to detect and track multiple types of agricultural machines regardless of whether they have GNSS capability. The system can identify machines through visual recognition of their physical characteristics, making it adaptable to both modern and older machinery. This multi-functional approach allows a single monitoring system to handle diverse equipment types without requiring different positioning technologies.
Solution Approach 2:
The patent substitutes GNSS-based positioning with vision-based positioning using imaging sensors. Instead of requiring each machine to have its own GNSS receiver for positioning, the system uses images captured by sensors to identify and locate machines based on their visual appearance. This substitution maintains positioning capability while dramatically improving adaptability to older machinery that lacks electronic positioning systems.
3Device complexity
If imaging sensors are used to detect and classify agricultural machines, then system complexity and cost are reduced, but measurement precision of machine identification may be affected
Solution Approach 1:
The patent applies preliminary action by using a detection model that is pre-trained and configured to recognize specific agricultural machines and their components. The system prepares classification categories and recognition algorithms in advance, enabling accurate identification when images are captured. This preliminary setup ensures that even though the system is simpler than GNSS-based solutions, it maintains adequate classification accuracy through pre-configured recognition capabilities.
Solution Approach 2:
The patent uses imaging sensors to create visual copies (images) of agricultural machines, which are then processed by detection models to extract identification information. This copying approach allows the system to analyze machine characteristics without physically interacting with the machines or requiring complex onboard electronics. The visual copies provide sufficient information for classification while keeping the overall system simple and cost-effective.
Data Source
AI summary
Systems and methods are provided for mapping one or more agricultural operations within a working environment. Image data is received from one or more imaging sensors mounted or otherwise associated with one or more agricultural machines within the working environment and configured to obtain image data indicative of the working environment and/or one or more objects located therein. The data is analysed utilising a detection model to classify one or more objects within the working environment to identify, from the one or more classified objects, one or more working machines within the working environment. Operational information for the working machine(s) is logged and/or updated in an operational map of the working environment in dependence on the identification of the classified object(s).