CN117523551A - Image processing system for evaluating and sorting durian quality - Google Patents

Image processing system for evaluating and sorting durian quality Download PDF

Info

Publication number
CN117523551A
CN117523551A CN202311571135.9A CN202311571135A CN117523551A CN 117523551 A CN117523551 A CN 117523551A CN 202311571135 A CN202311571135 A CN 202311571135A CN 117523551 A CN117523551 A CN 117523551A
Authority
CN
China
Prior art keywords
durian
image processing
weight
processing system
sorting
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202311571135.9A
Other languages
Chinese (zh)
Inventor
卡诺克波恩尼特尼蒂普鲁特
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Ka Nuokeboennitenidipulute
Original Assignee
Ka Nuokeboennitenidipulute
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Ka Nuokeboennitenidipulute filed Critical Ka Nuokeboennitenidipulute
Priority to CN202311571135.9A priority Critical patent/CN117523551A/en
Publication of CN117523551A publication Critical patent/CN117523551A/en
Pending legal-status Critical Current

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/60—Type of objects
    • G06V20/68—Food, e.g. fruit or vegetables
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/0464—Convolutional networks [CNN, ConvNet]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/0002—Inspection of images, e.g. flaw detection
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/20—Image preprocessing
    • G06V10/26—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/40—Extraction of image or video features
    • G06V10/44—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/764—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects
    • G06V10/765—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects using rules for classification or partitioning the feature space
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Multimedia (AREA)
  • Evolutionary Computation (AREA)
  • Health & Medical Sciences (AREA)
  • Artificial Intelligence (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Computing Systems (AREA)
  • Software Systems (AREA)
  • General Health & Medical Sciences (AREA)
  • Medical Informatics (AREA)
  • Databases & Information Systems (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • Computational Linguistics (AREA)
  • Data Mining & Analysis (AREA)
  • Molecular Biology (AREA)
  • General Engineering & Computer Science (AREA)
  • Mathematical Physics (AREA)
  • Quality & Reliability (AREA)
  • Investigating Materials By The Use Of Optical Means Adapted For Particular Applications (AREA)

Abstract

The invention discloses an image processing system for evaluating and sorting durian quality, which comprises an image acquisition system, an image processing system, a data training module, a hardware integration system, an object detection model and an action execution system, wherein the image acquisition system is used for acquiring durian quality of durian; the image acquisition system is provided with a high-resolution RGB camera, and the image processing system comprises a preprocessing unit, a target detection unit, a semantic segmentation unit and an action execution system; the data training module trains on the data set; the hardware integrated system comprises a camera system, a weight measurement system, a mechanical arm cabinet system, an internal network and a feedback mechanism; compared with the prior art, the invention has the advantages that: the present invention is designed to classify durian using advanced image processing techniques. By exploiting the capabilities of object detection and semantic segmentation algorithms, the system thoroughly analyzes the image of durian, facilitating classification in the following manner.

Description

Image processing system for evaluating and sorting durian quality
Technical Field
The invention relates to the technical field of image processing, in particular to an image processing system for evaluating and sorting durian quality.
Background
Durian praised as the king of fruit has unique quality. These characteristics, such as maturity, overall grade, and the presence of defects, are key to determining fruit market value and consumer demand. Historically, the assessment of these criteria has been dependent on human judgment, and this approach is fraught with subjectivity and inconsistency. Given the importance of durian in different markets, these inconsistencies can lead to different consumer experiences and can affect the perceived reliability of the seller.
The existing classification method mainly depends on manual work. Although manual inspectors have had many years of experience, the process is subjective in nature, resulting in potential negligence or discrepancy in quality assessment. Such inconsistencies can affect brand reputation, lead to potential revenue losses, and lead to supply chain inefficiency. Furthermore, manual sorting is very time consuming, limiting throughput, especially during the harvest season.
In view of these challenges, it is undeniable that a system is needed that not only automatically completes the sorting process, but also ensures accuracy, speed and consistency of durian quality assessment. Such a system would alleviate the challenges associated with manual sorting and set new industry benchmarks for reliability and efficiency.
Disclosure of Invention
The technical problem to be solved by the invention is to overcome the technical defects, and an image processing system for evaluating and sorting durian quality is specially designed for sorting durian by using advanced image processing technology. By exploiting the capabilities of object detection and semantic segmentation algorithms, the system thoroughly analyzes the image of durian, facilitating classification in the following manner.
In order to solve the problems, the technical scheme of the invention is as follows: the system comprises an image acquisition system, an image processing system, a data training module, a hardware integration system, an object detection model and an action execution system;
the image acquisition system is provided with a high-resolution RGB camera, can shoot durian at multiple angles when the durian moves along the conveyor system, is provided with an infrared sensor to detect internal defects of the durian, and provides additional data to improve the accuracy of defect detection;
the image processing system comprises a preprocessing unit, a target detection unit, a semantic segmentation unit and an action execution system;
the data training module trains on a data set, wherein the data set comprises more than 50,000 durian images collected in different seasons and regions, the images are annotated by experts, and the data is subjected to continuous model training through periodic expansion, so that the system is ensured to adapt to new durian varieties or defects which are not found before;
the hardware integrated system comprises a camera system, a weight measurement system, a mechanical arm cabinet system, an internal network and a feedback mechanism.
Further, the action execution system sorts durian, evaluates the quality of durian according to defects, maturity and grade by utilizing image processing, and comprises the following specific steps: the method comprises the steps of identifying and marking a region of interest on durian by adopting an object detection algorithm, providing a semantic segmentation technology, providing comprehensive analysis on the surface of durian, designing defects for identifying durian, distinguishing damage, scratch and decay of pests, determining the maturity of durian, distinguishing ripe fruits from immature fruits, grading durian according to the shape of durian, and ensuring consistency of quality assessment.
Further, the object detection algorithm is used for classifying the overall shape and outline of durian, and the popular object detection architecture utilized in the system comprises: YOLO (You Only Look Once) includes YOLOv3, YOLOv4, and YOLOv5; faster R-CNN: combining a Regional Proposal Network (RPN) with Fast R-CNN to realize efficient and accurate target detection; SSD (Single Shot MultiBox Detector): another real-time detection model focuses on the balance between speed and accuracy; retinaNet: with focus loss, retinaNet solves the class imbalance problem.
Further, the semantic segmentation technique identifies mature and immature parts of durian and distinguishes various types of diseases, and the semantic segmentation architecture that can be utilized by the semantic segmentation technique includes: deep Lab; U-Net: the architecture of the U-Net is suitable for various segmentation tasks due to the symmetrical extension paths; mask R-CNN: expanding the Faster R-CNN by adding a segmentation mask, thereby achieving object instance segmentation; PSPNet (pyramid scene parsing network): and collecting global context information by using a pyramid pool module, and enhancing a segmentation result.
Further, the preprocessing unit includes resizing, normalization, and addition (e.g., rotation and flipping) to ensure different data inputs and to maintain aspect ratios; the object detection unit classifies the shape and the visible maturity markers according to them; the semantic segmentation unit trains on the pixel annotation image of durian, marks defects and distinguishes mature parts from immature parts; the action execution system includes actuating a robotic arm to separate durian into designated bins or to mark them for further manual inspection. The system uses a real-time feedback loop to ensure fast response and minimal delay.
Further, the high resolution camera of the camera system with the infrared sensor captures detailed images of durian, providing visual and internal data to the processing system, the weight measurement system weighs each durian as it passes through the conveyor belt, the weight data is immediately transferred to the image processing system, the maturity can be determined from the weight and immature or overripe durian that does not meet the quality criteria is identified, the action execution system activates the robotic arm based on the decisions from the image processing system, the robotic arm sorts durian into designated bins based on the quality criteria, the hardware integration ensures a high speed internal network of rapid data exchange between the different components, ensures real time execution of image acquisition, weight measurement and sorting operations, minimizes delays and ensures efficiency, the sensors on the conveyor system provide feedback of durian position, weight and system action, ensures synchronization, and allows any necessary real time adjustment.
Further, a false negative minimization, dual camera system, continuous monitoring system, verification and test system is included.
Further, the method comprises the following steps:
(1) Defect classification
Capturing high resolution images of the durian surface under controlled illumination conditions by image analysis using an Image Acquisition System (IAS), object detection algorithms such as deep v3, YOLO and faster R-CNN process these images and highlight potential defect areas;
detailed analysis of highlighted areas by semantic segmentation models that distinguish various defect types based on unique patterns, discoloration or anomalies, pest lesions exhibiting obvious bite or perforation, bruises having irregular discoloration, and decay typically exhibiting pasty dark areas;
(2) Expiration ordering
Analyzing the color spectrum and texture of the durian peel by visual indicators, the object detection model evaluating parameters like vibration of the stems, skin gloss and any visual cues that indicate the doneness of durian;
when durian passes through the conveyor belt, a series of high-precision weighing sensors integrated in the structure of the conveyor belt are used for detecting tiny weight changes by adopting a strain gauge technology, a data acquisition system interprets the analog weight signals, converts the analog weight signals into a digital format and sends the digital weight signals to a main processing unit for analysis, and the weight of the durian is related to the water content and the pulp density of the durian;
surface color variations are distinguished by detailed pixel semantic segmentation, specific hues or gradients are associated with maturity levels, and a composite maturity score is assigned.
(3) Class classification
Using the captured images, advanced contour detection algorithms determine the geometric contour of durian, evaluate parameters such as sphericity, symmetry and surface smoothness, and any deviation from an ideal shape, such as bumps, flat spots or asymmetry, will be quantified and included in the grading criteria.
Weight considerations: weight is a secondary grading criterion, and the optimal weight range for quality durian is set as such. Those that clearly fall outside this range, indicating potential under-or over-ripening, are ranked accordingly;
(4) Weight-based packaging recommendations:
the conveyor belt integrates high-resolution weight measurement modules at specified intervals, the modules adopt capacitive load sensors, rapid and accurate readings can be provided, sensor data are interfaced with a special microcontroller, the microcontroller constantly calibrates and adjusts factors such as temperature drift or vibration of the conveyor belt, and a filtering algorithm measures for a plurality of times on average so as to eliminate transient effects and ensure highly accurate weight readings;
based on weight classification, the system recommends packaging types, combines similar weights together to achieve consistent packaging, and the algorithm considers the weight and size of durian to suggest optimal packaging materials, such as specific box sizes or protective cushioning, to ensure safe transportation and storage.
Compared with the prior art, the invention has the advantages that:
1. the invention adopts an object detection and semantic segmentation algorithm to analyze durian images and classifies the durian images according to the following conditions: defects: distinguishing whether durian has internal and external defects such as insect damage, bruise and rot; maturity degree: distinguishing between ripe and immature fruits; grade: classifying durian according to shape can provide a consistent, rapid and efficient method of classifying durian, reducing human error and improving productivity.
2. The invention eliminates subjectivity of manual sorting, ensures consistency of quality evaluation, accelerates sorting process, thereby improving throughput and reducing operation cost, adopts extensive algorithm terms for design, and allows potential upgrading or integration with updating technology in future.
Drawings
Fig. 1 is a flow chart of an image processing system for durian quality assessment and sorting of the present invention.
Detailed Description
In order to make the contents of the present invention more clearly understood, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention.
Example 1
The image processing system for evaluating and sorting the durian quality comprises an image acquisition system, an image processing system, a data training module, a hardware integration system, an object detection model and an action execution system;
the image acquisition system is provided with a high-resolution RGB camera, so that durian can be shot at multiple angles when moving along the conveyor belt system, an infrared sensor is arranged to detect internal defects of durian, and extra data is provided to improve the accuracy of defect detection;
the image processing system comprises a preprocessing unit, a target detection unit, a semantic segmentation unit and an action execution system;
the data training module trains on a data set, wherein the data set comprises more than 50,000 durian images collected in different seasons and regions, the images are annotated by experts, and the data is subjected to continuous model training through periodic expansion, so that the system is ensured to adapt to new durian varieties or defects which are not found before;
the hardware integrated system comprises a camera system, a weight measurement system, a mechanical arm cabinet system, an internal network and a feedback mechanism.
The action execution system sorts durian, and evaluates the quality of durian according to the defects, maturity and grade by utilizing image processing, and the specific steps are as follows: the method comprises the steps of identifying and marking a region of interest on durian by adopting an object detection algorithm, providing a semantic segmentation technology, providing comprehensive analysis on the surface of durian, designing defects for identifying durian, distinguishing damage, scratch and decay of pests, determining the maturity of durian, distinguishing ripe fruits from immature fruits, grading durian according to the shape of durian, and ensuring consistency of quality assessment.
The object detection algorithm is used for classifying the overall shape and outline of durian, and popular object detection architecture utilized in the system includes: YOLO (You Only Look Once) includes YOLOv3, YOLOv4, and YOLOv5; faster R-CNN: combining a Regional Proposal Network (RPN) with Fast R-CNN to realize efficient and accurate target detection; SSD (Single Shot MultiBox Detector): another real-time detection model focuses on the balance between speed and accuracy; retinaNet: with focus loss, retinaNet solves the class imbalance problem.
Semantic segmentation techniques recognize the mature and immature parts of durian and distinguish various types of diseases, semantic segmentation architecture that can be exploited by semantic segmentation techniques includes: deep Lab; U-Net: the architecture of the U-Net is suitable for various segmentation tasks due to the symmetrical extension paths; mask R-CNN: expanding the Faster R-CNN by adding a segmentation mask, thereby achieving object instance segmentation; PSPNet (pyramid scene parsing network): and collecting global context information by using a pyramid pool module, and enhancing a segmentation result.
The preprocessing unit includes resizing, normalization, and adding (e.g., rotation and flipping) to ensure different data inputs and to maintain aspect ratios; the object detection unit classifies the shape and the visible maturity markers according to them; the semantic segmentation unit trains on the pixel annotation image of durian, marks defects and distinguishes a mature part from an immature part; the action execution system includes actuating a robotic arm to separate durian into designated bins or to mark them for further manual inspection. The system uses a real-time feedback loop to ensure fast response and minimal delay.
The camera system captures detailed images of durian with a high resolution camera equipped with an infrared sensor, provides visual and internal data to the processing system, the weight measurement system weighs each durian as it passes through the conveyor belt, the weight data is immediately transferred to the image processing system, the maturity can be determined from the weight and unripe or overripe durian that does not meet the quality criteria can be identified, the action execution system activates the robotic arm based on the decisions from the image processing system, the robotic arm sorts durian into designated bins based on the quality criteria, the hardware integration ensures a high speed internal network of rapid data exchange between the different components, ensures real time execution of image acquisition, weight measurement and sorting operations, minimizes delays and ensures efficiency, the sensors on the conveyor system provide feedback of position, weight and system action of durian,
ensure synchronization and allow any necessary real-time adjustment.
Safety and reliability measures include false negative minimization, the system is designed to preferentially minimize false negatives, especially for critical defects like rot, by setting a conservative threshold, ensuring that problematic durian is marked for manual inspection.
The dual camera system ensures that even if one camera misses a defect, the other can capture. The integrated dashboard allows operators to oversee the decisions of the system in real time and intervene if necessary.
Prior to commercial deployment, the system underwent rigorous testing. Three main durian processing factories perform trial points for six months, more than 100 thousands of durian are processed, and the system realizes the accuracy of 98.5 percent, which is far more than manual sorting.
Example two
An image processing system for durian quality assessment and sorting comprising the steps of:
(1) Defect classification
Capturing high resolution images of the durian surface under controlled illumination conditions by image analysis using an Image Acquisition System (IAS), object detection algorithms such as deep v3, YOLO and faster R-CNN process these images and highlight potential defect areas;
detailed analysis of highlighted areas by semantic segmentation models that distinguish various defect types based on unique patterns, discoloration or anomalies, pest lesions exhibiting obvious bite or perforation, bruises having irregular discoloration, and decay typically exhibiting pasty dark areas;
(2) Expiration ordering
Analyzing the color spectrum and texture of the durian peel by visual indicators, the object detection model evaluating parameters like vibration of the stems, skin gloss and any visual cues that indicate the doneness of durian;
when durian passes through the conveyor belt, a series of high-precision weighing sensors integrated in the structure of the conveyor belt are used for detecting tiny weight changes by adopting a strain gauge technology, a data acquisition system interprets the analog weight signals, converts the analog weight signals into a digital format and sends the digital weight signals to a main processing unit for analysis, and the weight of the durian is related to the water content and the pulp density of the durian;
surface color variations are distinguished by detailed pixel semantic segmentation, specific hues or gradients are associated with maturity levels, and a composite maturity score is assigned.
(3) Class classification
Using the captured images, advanced contour detection algorithms determine the geometric contour of durian, evaluate parameters such as sphericity, symmetry and surface smoothness, and any deviation from an ideal shape, such as bumps, flat spots or asymmetry, will be quantified and included in the grading criteria.
Weight considerations: weight is a secondary grading criterion, and the optimal weight range for quality durian is set as such. Those that clearly fall outside this range, indicating potential under-or over-ripening, are ranked accordingly;
(4) Weight-based packaging recommendations:
the conveyor belt integrates high-resolution weight measurement modules at specified intervals, the modules adopt capacitive load sensors, rapid and accurate readings can be provided, sensor data are interfaced with a special microcontroller, the microcontroller constantly calibrates and adjusts factors such as temperature drift or vibration of the conveyor belt, and a filtering algorithm measures for a plurality of times on average so as to eliminate transient effects and ensure highly accurate weight readings;
based on weight classification, the system recommends packaging types, combines similar weights together to achieve consistent packaging, and the algorithm considers the weight and size of durian to suggest optimal packaging materials, such as specific box sizes or protective cushioning, to ensure safe transportation and storage.
The invention and its embodiments have been described above without limitation. In summary, if one of ordinary skill in the art is informed by this disclosure, a structural manner and an embodiment similar to the technical solution should not be creatively devised without departing from the gist of the present invention.

Claims (8)

1. An image processing system for evaluating and sorting durian quality, characterized in that: the system comprises an image acquisition system, an image processing system, a data training module, a hardware integration system, an object detection model and an action execution system;
the image acquisition system is provided with a high-resolution RGB camera, can shoot durian at multiple angles when the durian moves along the conveyor system, is provided with an infrared sensor to detect internal defects of the durian, and provides additional data to improve the accuracy of defect detection;
the image processing system comprises a preprocessing unit, a target detection unit, a semantic segmentation unit and an action execution system;
the data training module trains on a data set, wherein the data set comprises more than 50,000 durian images collected in different seasons and regions, the images are annotated by experts, and the data is subjected to continuous model training through periodic expansion, so that the system is ensured to adapt to new durian varieties or defects which are not found before;
the hardware integrated system comprises a camera system, a weight measurement system, a mechanical arm cabinet system, an internal network and a feedback mechanism.
2. The image processing system for durian quality assessment and sorting of claim 1, wherein: the action execution system sorts durian, evaluates the quality of durian according to the defect, maturity and grade by utilizing image processing, and comprises the following specific steps: the method comprises the steps of identifying and marking a region of interest on durian by adopting an object detection algorithm, providing a semantic segmentation technology, providing comprehensive analysis on the surface of durian, designing defects for identifying durian, distinguishing damage, scratch and decay of pests, determining the maturity of durian, distinguishing ripe fruits from immature fruits, grading durian according to the shape of durian, and ensuring consistency of quality assessment.
3. The image processing system for durian quality assessment and sorting according to claim 2, characterized in that: the object detection algorithm is used for classifying the overall shape and outline of durian, and the popular object detection architecture utilized in the system comprises: YOLO (You Only Look Once) includes YOLOv3, YOLOv4, and YOLOv5; faster R-CNN: combining a Regional Proposal Network (RPN) with Fast R-CNN to realize efficient and accurate target detection; SSD (Single Shot MultiBox Detector): another real-time detection model focuses on the balance between speed and accuracy; retinaNet: with focus loss, retinaNet solves the class imbalance problem.
4. The image processing system for durian quality assessment and sorting of claim 1, wherein: the semantic segmentation technique identifies mature and immature parts of durian and distinguishes various types of diseases, and the semantic segmentation architecture that can be utilized by the semantic segmentation technique includes: deep Lab; U-Net: the architecture of the U-Net is suitable for various segmentation tasks due to the symmetrical extension paths; mask R-CNN: expanding the Faster R-CNN by adding a segmentation mask, thereby achieving object instance segmentation; PSPNet (pyramid scene parsing network): and collecting global context information by using a pyramid pool module, and enhancing a segmentation result.
5. The image processing system for durian quality assessment and sorting of claim 1, wherein: the preprocessing unit includes resizing, normalization, and adding (e.g., rotation and flipping) to ensure different data inputs and to maintain aspect ratios; the object detection unit classifies the shape and the visible maturity markers according to them; the semantic segmentation unit trains on the pixel annotation image of durian, marks defects and distinguishes mature parts from immature parts; the action execution system includes actuating a robotic arm to separate durian into designated bins or to mark them for further manual inspection. The system uses a real-time feedback loop to ensure fast response and minimal delay.
6. The image processing system for durian quality assessment and sorting of claim 1, wherein: the camera system captures detailed images of durian with a high resolution camera equipped with an infrared sensor providing visual and internal data to the processing system, the weight measurement system weighs each durian as it passes by the conveyor belt, the weight data is immediately transferred to the image processing system, the maturity can be determined from the weight and immature or overripe durian that does not meet the quality criteria is identified, the action execution system activates a robotic arm based on the decisions from the image processing system, the robotic arm sorts durian into specified boxes based on the quality criteria, the hardware integration ensures a high speed internal network of rapid data exchange between the different components, real time execution of image acquisition, weight measurement and sorting operations is ensured, delay is minimized and efficiency is ensured, the sensors on the conveyor system provide feedback of durian position, weight and system action, synchronization is ensured, and any necessary real time adjustment is allowed.
7. The image processing system for durian quality assessment and sorting of claim 1, wherein: the system also comprises a false negative minimization, a double-camera system, a continuous monitoring system and a verification and test system.
8. The image processing system for durian quality assessment and sorting of claim 1, wherein: the method comprises the following steps:
(1) Defect classification
Capturing high resolution images of the durian surface under controlled illumination conditions by image analysis using an Image Acquisition System (IAS), object detection algorithms such as deep v3, YOLO and faster R-CNN process these images and highlight potential defect areas;
detailed analysis of highlighted areas by semantic segmentation models that distinguish various defect types based on unique patterns, discoloration or anomalies, pest lesions exhibiting obvious bite or perforation, bruises having irregular discoloration, and decay typically exhibiting pasty dark areas;
(2) Expiration ordering
Analyzing the color spectrum and texture of the durian peel by visual indicators, the object detection model evaluating parameters like vibration of the stems, skin gloss and any visual cues that indicate the doneness of durian;
when durian passes through the conveyor belt, a series of high-precision weighing sensors integrated in the structure of the conveyor belt are used for detecting tiny weight changes by adopting a strain gauge technology, a data acquisition system interprets the analog weight signals, converts the analog weight signals into a digital format and sends the digital weight signals to a main processing unit for analysis, and the weight of the durian is related to the water content and the pulp density of the durian;
surface color variations are distinguished by detailed pixel semantic segmentation, specific hues or gradients are associated with maturity levels, and a composite maturity score is assigned.
(3) Class classification
Using the captured images, advanced contour detection algorithms determine the geometric contour of durian, evaluate parameters such as sphericity, symmetry and surface smoothness, and any deviation from an ideal shape, such as bumps, flat spots or asymmetry, will be quantified and included in the grading criteria.
Weight considerations: weight is a secondary grading criterion, and the optimal weight range for quality durian is set as such. Those that clearly fall outside this range, indicating potential under-or over-ripening, are ranked accordingly;
(4) Weight-based packaging recommendations:
the conveyor belt integrates high-resolution weight measurement modules at specified intervals, the modules adopt capacitive load sensors, rapid and accurate readings can be provided, sensor data are interfaced with a special microcontroller, the microcontroller constantly calibrates and adjusts factors such as temperature drift or vibration of the conveyor belt, and a filtering algorithm measures for a plurality of times on average so as to eliminate transient effects and ensure highly accurate weight readings;
based on weight classification, the system recommends packaging types, combines similar weights together to achieve consistent packaging, and the algorithm considers the weight and size of durian to suggest optimal packaging materials, such as specific box sizes or protective cushioning, to ensure safe transportation and storage.
CN202311571135.9A 2023-11-23 2023-11-23 Image processing system for evaluating and sorting durian quality Pending CN117523551A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202311571135.9A CN117523551A (en) 2023-11-23 2023-11-23 Image processing system for evaluating and sorting durian quality

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202311571135.9A CN117523551A (en) 2023-11-23 2023-11-23 Image processing system for evaluating and sorting durian quality

Publications (1)

Publication Number Publication Date
CN117523551A true CN117523551A (en) 2024-02-06

Family

ID=89751035

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202311571135.9A Pending CN117523551A (en) 2023-11-23 2023-11-23 Image processing system for evaluating and sorting durian quality

Country Status (1)

Country Link
CN (1) CN117523551A (en)

Similar Documents

Publication Publication Date Title
Makky et al. Development of an automatic grading machine for oil palm fresh fruits bunches (FFBs) based on machine vision
Ireri et al. A computer vision system for defect discrimination and grading in tomatoes using machine learning and image processing
Khojastehnazhand et al. Development of a lemon sorting system based on color and size
Kondo Automation on fruit and vegetable grading system and food traceability
Baigvand et al. Machine vision system for grading of dried figs
Laykin et al. Image–processing algorithms for tomato classification
Omid et al. An expert egg grading system based on machine vision and artificial intelligence techniques
Blasco et al. Machine vision system for automatic quality grading of fruit
US6271520B1 (en) Item defect detection apparatus and method
Noordam et al. High-speed potato grading and quality inspection based on a color vision system
Jarimopas et al. An experimental machine vision system for sorting sweet tamarind
Rokunuzzaman et al. Development of a low cost machine vision system for sorting of tomatoes.
Deng et al. A carrot sorting system using machine vision technique
Heinemann et al. Machine vision inspection of ‘Golden Delicious’ apples
Ali et al. Automated fruit grading system
Vakilian et al. An apple grading system according to European fruit quality standards using Gabor filter and artificial neural networks
JP7248317B2 (en) Tofu production system
Arango et al. Quality control of apples by means of convolutional neural networks-Comparison of bruise detection by color images and near-infrared images
Raut et al. Assessment of fruit maturity using digital image processing
Thong et al. Mango classification system uses image processing technology and artificial intelligence
Bautista et al. Automated sorter and grading of tomatoes using image analysis and deep learning techniques
Mohamed et al. Development of a real-time machine vision prototype to detect external defects in some agricultural products
Ahmad et al. Development of automatic grading machine prototype for citrus using image processing
Esehaghbeygi et al. Digital image processing for quality ranking of saffron peach
Thong et al. Mango sorting mechanical system combines image processing

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination