Pallet Tracking via Neural Network Image Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems face challenges in accurately tracking pallets through transportation paths due to obscured pallets and inclusion of non-target pallets in images, making it difficult to determine the number of pallets and detect additions or removals between nodes.
Innovation Solution
A computer-implemented method using an artificial neural network to process images from various imaging devices, filter out non-target pallets, and adjust weights to improve accuracy in counting and matching pallets, allowing for the tracking of pallets from a specific source across a transportation path.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If various imaging devices are used to gather images of pallet shipments, then more comprehensive tracking data can be obtained, but it becomes more difficult to accurately determine the number of pallets due to obscured pallets and inclusion of non-target pallets
Solution Approach 1:
The system segments the pallet tracking problem by using multiple imaging devices positioned at different locations and angles along the transportation path. Each device captures images of specific sections or perspectives of the pallet shipment, allowing the system to piece together complete information while maintaining accuracy through distributed observation points.
Solution Approach 2:
The patent introduces an intermediary processing system that receives images from multiple imaging devices and applies image processing techniques to identify, filter, and count pallets. This intermediary layer reconciles the conflicting requirements by processing the raw image data to eliminate obscured views and non-target pallets, producing accurate count information.
2Productivity
If images are captured at multiple nodes along the transportation path, then pallet movement can be tracked, but it becomes difficult to determine whether pallets have been added or removed between nodes
Solution Approach 1:
The system implements feedback mechanisms by comparing pallet counts and identifiers from images captured at different nodes along the transportation path. The processing system analyzes sequences of images to detect changes in pallet composition, providing feedback on whether pallets have been added or removed between monitoring points, thereby ensuring reliable tracking.
Solution Approach 2:
The patent applies preliminary action by establishing baseline pallet inventories at the起始 node before transportation begins. This preliminary documentation of pallet identifiers, quantities, and positions enables subsequent comparison at intermediate and final nodes to detect any additions or removals during transit.
3Device complexity
If traditional image processing methods are used to count pallets, then the system is simpler to implement, but accuracy is reduced due to obscured pallets and non-target pallets in images
Solution Approach 1:
The patent introduces an intermediary artificial neural network processing layer between the imaging devices and the final counting output. This neural network intermediary automatically filters out non-target pallets, handles obscured views through pattern recognition, and provides accurate pallet identification without requiring complex manual image processing procedures.
Solution Approach 2:
The system replaces traditional mechanical or manual image processing methods with an artificial neural network-based automated recognition system. This substitution maintains relative system simplicity while dramatically improving accuracy by using machine learning algorithms to identify and count pallets even when partially obscured or when non-target pallets are present in the field of view.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Provided are methods, devices, and computer-program products for tracking goods carriers from a particular source. According to some embodiments of the invention, a computer-implemented method includes training an artificial neural network to count the number of goods carriers from a particular source within an image. Further, the method includes receiving a first image file generated by a first imaging device; using the trained artificial neural network to determine a first number of goods carriers from the particular source in the first image; receiving a second image file generated by a second imaging device; using the trained artificial neural network to determine a second number of goods carriers from the particular source in the second image; and determining whether the first number of goods carriers from the particular source in the first image is equal to the second number of goods carriers from the particular source in the second image.