Multi-Band Produce Inspection for Internal Defect Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing NIR technologies face challenges in accurately detecting small and localized internal defects in agricultural produce, such as vascular browning in apples and high false positive rates in onions, making them impractical for effective sorting and grading.
Innovation Solution
A method involving simultaneous use of low and high band light sources to generate spatial profiles of spectroscopic values, which are compared to reference profiles to determine internal quality attributes, utilizing a system with laser diodes, photodiodes, and computational analysis to assess defects like botrytis fungus and pseudomonas bacteria in onions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional NIR spectroscopy is used for detecting internal defects, then the method is simple and non-destructive, but the detection precision for small and localized defects is poor
Solution Approach 1:
The spectral range is segmented into multiple bands (e.g., visible, near-infrared, short-wave infrared) with different wavelength ranges. Each band provides complementary information for detecting different types of defects, enabling precise detection of small and localized internal defects while maintaining system manageability through modular architecture
Solution Approach 2:
The system transitions from single-wavelength or narrow-band measurement to multi-dimensional spectral analysis by measuring across multiple wavelength bands simultaneously. This dimensional expansion enables detection of defects that are invisible to single-band methods, significantly improving detection precision without proportionally increasing system complexity
2Measurement precision
If single ratio of two signals is used for defect detection, then the method is simple, but the detection capability for vascular browning is insufficient
Solution Approach 1:
The spectral measurement is divided into multiple discrete wavelength bands rather than using a single ratio. Each band provides specific spectral information, and the combination of multiple ratios creates a comprehensive spectral signature that enables detection of subtle defects like vascular browning that single ratios cannot detect
Solution Approach 2:
The system creates a composite spectral signature by combining measurements from multiple wavelength bands. This composite approach integrates information from different spectral regions, enhancing the detection capability for complex defects like vascular browning while maintaining measurement simplicity through automated multi-parameter analysis
3Measurement precision
If multiple measurements from different locations are taken, then the measurement precision improves, but the productivity decreases due to time consumption
Solution Approach 1:
The system performs continuous multi-point spectral measurement across the entire surface of the agricultural product without interruption. Multiple measurements are taken simultaneously at different locations during a single pass through the inspection system, maintaining continuous productive action while achieving high precision through comprehensive spatial sampling
Solution Approach 2:
The system pre-establishes reference spectral profiles and spatial measurement patterns before actual inspection. By preparing measurement protocols and reference data in advance, the system can rapidly execute multiple measurements at different locations without time-consuming setup or analysis during the grading process, thus maintaining high productivity
4Reliability
If conventional NIR technology is used for onion inspection, then the method is simple, but the reliability is poor with high false positive rates
Solution Approach 1:
The spectral analysis is divided into multiple discrete wavelength bands, each providing specific diagnostic information. By segmenting the spectral range and analyzing each band separately before integration, the system reduces false positives through more targeted detection while maintaining manageable system complexity through automated multi-parameter evaluation
Solution Approach 2:
The system incorporates reference spectral profiles and real-time comparative analysis to feedback-adjust defect detection. By continuously comparing measured spectra against established references and using this feedback to refine classification decisions, the system significantly improves reliability and reduces false positives while maintaining operational simplicity through automated decision-making
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 method provides accurate and efficient detection of internal defects in agricultural produce, reducing false positives and enabling effective sorting and grading at high speeds.
Implementation Method 1
receiving a plurality of first spectroscopic values obtained from directing low band light in a first wavelength associated to a low band of wavelengths from at least one low band light source at least partly through the article toward at least one detector
Data Source
AI summary
An aspect of the invention provides a method for determining at least one internal quality attribute of an article (102) of agricultural produce. The method includes receiving a plurality of first spectroscopic values obtained from directing low band light in a first wavelength associated to a low band of wavelengths from at least one low band light source (104) at least partly through the article (102) toward at least one detector (120); receiving a plurality of second spectroscopic values obtained from directing high band light in a second wavelength associated to a high band of wavelengths from at least one high band light source (106) at least partly through the article (102) toward the at least one detector (120); determining at least one measured spatial profile associated to the article, the at least one measured spatial profile comprising at least one of a plurality of ratios of respective first spectroscopic values to respective second spectroscopic values, a plurality of ratios of respective second spectroscopic values to respective first spectroscopic values; and determining the at least one internal quality attribute at least partly from a comparison of the at least one measured spatial profile with at least one reference spatial profile associated to a class of articles of agricultural produce.