Poultry Leg Bone Localization Using Optical Imaging and AI
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for locating bony parts in poultry legs require X-ray imaging, which is costly, maintenance-intensive, and necessitates skilled personnel, posing safety challenges.
Innovation Solution
A method involving optical imaging and X-ray imaging during training, followed by optical-only location using a neural network, with defined reference points for bony parts, allowing precise identification without X-ray imaging during operation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If X-ray imaging is used to locate bony parts in poultry legs, then measurement precision is improved, but device complexity and safety risks increase
Solution Approach 1:
The patent applies preliminary action by using X-ray imaging only during the training phase to create reference data and train the neural network. Once trained, the system operates using only optical imaging, eliminating the need for X-ray equipment during actual production. This resolves the contradiction by concentrating the measurement precision requirement into a one-time training process rather than continuous operation.
Solution Approach 2:
The patent creates a virtual copy of the X-ray reference data by training a neural network to recognize bony part locations from optical images. The neural network learns to map optical image features to bony part positions based on X-ray trained reference points, effectively copying the localization capability without requiring physical X-ray equipment during operation.
2Reliability
If X-ray imaging system is deployed for continuous operation, then bony part detection reliability is improved, but operational costs and maintenance requirements increase
Solution Approach 1:
The system performs preliminary X-ray imaging and reference point definition during the training phase to establish reliable detection patterns. The neural network learns from these X-ray reference data and subsequently provides reliable detection using only optical imaging, eliminating ongoing X-ray equipment costs and maintenance requirements while preserving detection reliability.
Solution Approach 2:
The patent replaces the mechanical X-ray imaging system with an optical imaging system for continuous operation. The neural network, trained on X-ray reference data, substitutes the physical X-ray measurement process with computational analysis of optical images, dramatically reducing equipment costs and maintenance needs while maintaining detection reliability.
3Measurement precision
If skilled personnel are required to handle X-rays, then measurement precision is improved, but ease of operation deteriorates
Solution Approach 1:
The patent performs the complex task of X-ray image analysis and reference point definition during the training phase, where skilled personnel can work. The trained neural network then automatically performs bony part location without requiring skilled personnel during operation, improving ease of operation while preserving measurement precision through the trained model.
Solution Approach 2:
The neural network becomes self-sufficient after training, automatically locating bony parts in optical images without requiring skilled personnel intervention during operation. The system serves itself by using the trained model to process images and provide accurate bony part locations, eliminating the need for specialized human operators while maintaining precision.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
Enables precise bony part location in poultry legs without X-ray imaging, reducing costs and safety risks while maintaining precision, facilitating fully automatic deboning.
Implementation Method 1
recording images of the front sides or back sides of the poultry legs in the optically visible wavelength range by means of an optical camera
Implementation Method 2
irradiating the back side or front side of the poultry legs with X-rays of an X-ray source and recording X-ray images on the side of the poultry legs that is remote from the X-ray source
Data Source
AI summary
A method, arrangement and apparatus are provided for training at least one neural network for locating bony parts present in a poultry leg. The method includes the steps: conveying the poultry legs in a conveying direction by a conveying device; acquiring digital images of the front side or back side of each of the poultry legs conveyed past the imaging system by a first optical imaging system; sequentially providing the digital images as input data to a first neural network configured for locating the bony parts, wherein the first neural network for locating the bony parts has been trained by a described method; and determining position data of the bony parts by the first neural network; and providing the position data, for display and/or transmission, to a downstream machine for processing the poultry legs on the basis of the determined position data.