Field Obstruction Detection in Farming Machines Using Unified CNNs
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Farming machines face challenges in recognizing a diverse range of obstructions in fields, leading to potential collisions and unsafe operating environments, as current systems require extensive data collection and manual labeling for training models, which is time-consuming and inefficient.
Innovation Solution
A farming machine equipped with sensors that capture image data and employs a convolutional neural network obstruction identification module trained on images of both common and uncommon obstructions, allowing it to autonomously identify and navigate around obstacles by modifying its treatment instructions in real-time.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If traditional multiple models are used to detect obstructions, then detection coverage is improved, but data collection time and manual labeling effort increase significantly
Solution Approach 1:
The patent employs a single convolutional neural network model that performs multiple functions: detecting various obstruction types (rocks, hay bales, workers, animals), classifying them by category, and providing location information. This universal model replaces the traditional approach of using separate specialized models for each obstruction type, thereby maintaining comprehensive detection coverage while eliminating the need to collect and label separate datasets for each category.
Solution Approach 2:
The patent changes the training parameters and architecture of the neural network model to handle diverse obstruction types within a single model. By adjusting the model's parameters, training data composition, and classification categories during training, the system achieves versatile detection capabilities without requiring multiple separate models, thus reducing data collection and labeling requirements.
2Reliability
If more obstruction types are detected, then safety is improved, but the complexity of the detection system increases
Solution Approach 1:
The patent merges multiple detection functions into a single integrated neural network model. Instead of having separate models for detecting different obstruction types, the system combines all detection, classification, and identification functions into one unified model, reducing system complexity while maintaining the ability to detect multiple obstruction types for improved safety.
Solution Approach 2:
The neural network model is designed to autonomously perform classification and identification of various obstruction types without requiring complex external processing systems. The model self-adapts to different obstruction categories through training, eliminating the need for complex rule-based systems or multiple specialized detectors, thereby reducing overall system complexity while enhancing safety.
3Measurement precision
If manual labeling of obstruction images is performed, then model training accuracy is improved, but labor requirements and costs increase
Solution Approach 1:
The system employs automated image labeling techniques where the neural network performs self-labeling or semi-automated labeling of obstruction images during the training process. This reduces or eliminates the need for manual human labeling while maintaining model training accuracy, thereby decreasing labor requirements and costs associated with data preparation.
Data Source
AI summary
A control system of a farming machine is configured to identify obstruction in a field from image data of the field. The control system accesses an obstruction model configured to identify obstructions in a field from image data of the field. The obstruction model is generated by accessing image data of obstructions in a training field, each obstruction corresponding to a prescribed action occurring at a prescribed time, labelling the image data of the obstructions, and training the obstruction model based on the labelled image data. The control system captures image data of the field including an obstruction and inputs the image data into the obstruction model. Responsive to identifying the obstruction in the field, the control system modifies treatment instructions of the farming machine such that the farming machine performs a implements a farming objective while avoiding the obstruction in the field.