X-ray Inspection Device for Overlapping Object Counting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing X-ray inspection devices cannot accurately count the number of objects with a predetermined shape that are randomly arranged and overlapping in a package, as they fail to distinguish individual objects in transmission images.
Innovation Solution
An inspection device equipped with a light emitter and detector, along with a controller that specifies object regions based on pixel occupancy and calculates the total number of objects by estimating mass per object using grayscale information, allowing for separation of overlapping objects and accurate counting.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional X-ray inspection devices are used to inspect randomly arranged objects, then the inspection process is simple, but the devices cannot accurately count the number of objects because they appear overlapping in transmission images
Solution Approach 1:
The controller segments the transmission image into multiple candidate regions based on pixel occupancy thresholds. First, a first candidate region is extracted where pixel occupancy exceeds a first threshold, indicating potential object presence. Then, a second candidate region is extracted where pixel occupancy exceeds a second (lower) threshold, capturing additional objects. This segmentation approach divides the complex task of counting overlapping objects into manageable regional analysis steps, enabling accurate object counting while maintaining reasonable processing complexity.
2Measurement precision
If the inspection device attempts to separate overlapping objects to count them individually, then counting accuracy improves, but processing time increases significantly
Solution Approach 1:
Instead of attempting to fully separate all overlapping objects which would be time-consuming, the device performs partial separation by extracting candidate regions based on pixel occupancy thresholds. This partial action approach extracts sufficient information to count objects accurately without requiring complete separation of all overlapping objects, thus reducing processing time while maintaining counting precision.
Solution Approach 2:
The device uses two different pixel occupancy thresholds (first threshold and second threshold) to extract different candidate regions. By changing the threshold parameter, the system can efficiently capture objects at different visibility levels in the transmission image, enabling fast processing while maintaining accurate object detection and counting.
3Productivity
If the device uses pixel occupancy thresholds to extract candidate regions, then object detection speed improves, but accuracy may decrease for objects with similar pixel occupancy
Solution Approach 1:
The device moves from a single-threshold approach to a two-dimensional threshold space by using both a first threshold and a second threshold to extract different candidate regions. This dimensional expansion allows the system to capture objects that might be missed by a single threshold, improving detection accuracy while maintaining the speed benefits of threshold-based processing. The multi-threshold approach effectively adds a layer of discrimination capability.
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
The device accurately calculates the number of objects in packages with randomly arranged and overlapping items, improving processing speed and accuracy even when object masses vary.
Implementation Method 1
a light emitter that emits light onto a product in which a plurality of objects each having a predetermined shape are contained, a light detector that detects light transmitted through the product
Data Source
Figure 1
Figure 2
Figure 3
AI summary
An inspection device is provided that is capable of counting the number of contained objects of a product in which objects each having a predetermined shape are contained in a package in an overlapping manner. An inspection device 1 includes a total-number calculator 11b that estimates a total mass of a plurality of objects based on grayscale information on a transmission image and calculates the total number of objects by dividing the total mass estimated by a mass of one object.