Threshold Segmentation Algorithm Research Articles

Quantitative detection and sorting of broken kernels and chalky grains in milled rice using computer vision algorithms

The milled rice industry faces challenges in effectively detecting and sorting broken kernels and chalky grains. This paper presents a computer vision approach to detect and characterize broken kernels and chalky grains in three milled rice varieties (HM105, WYD4, and JH6). Broken kernel sorting was validated against an ISO-derived recognition method, while the identification of chalky grains was validated using a DSC to assess thermal behavior. The logistic regression showed a strong discriminatory ability in identifying broken kernels. Unlike the ISO method, logistic regression achieved high sorting accuracy for HM105, WYD4, and JH6. The iterative threshold segmentation algorithm also accurately segmented chalky regions in milled rice grains. Regression analysis revealed a significant positive correlation (R2 ≥ 0.94) between the chalky region area and ΔH values. Therefore, this study introduces a robust and rapid methodology for quantitatively detecting milled rice, with the potential for automation and intelligentization in rice-based food processes.

Online Measurement of Melt-Pool Width in Direct Laser Deposition Process Based on Binocular Vision and Perspective Transformation.

Direct laser deposition (DLD) requires high-energy input and causes poor stability and portability. To improve the deposited layer quality, conducting online measurements and feedback control of the dimensions, temperature, and other melt-pool parameters during deposition is essential. Currently, melt-pool dimension measurement is mainly based on machine vision methods, which can mostly detect only the deposition direction of a single melt pool, limiting their measurement range and applicability. We propose a binocular-vision-based online measurement method to detect the melt-pool width during DLD. The method uses a perspective transformation algorithm to align multicamera measurements into a single-coordinate system and a fuzzy entropy threshold segmentation algorithm to extract the melt-pool true contour. This effectively captures melt-pool width information in various deposition directions. A DLD measurement system was constructed, establishing an online model that maps the melt-pool width to the offline deposited layer width, validating the accuracy of the binocular vision system in measuring melt-pool width at different deposition angles. The method achieved high accuracy for melt-pool measurements within certain deposition angle ranges. Within the 30°-60° measurement range, the average error is 0.056 mm, with <3% error. The proposed method enhances the detectable range of melt-pool widths, improving cladding layers and parts.

Online Measurement of Melt-Pool Width in Direct Laser Deposition Process Based on Binocular Vision and Perspective Transformation

Direct laser deposition (DLD) requires high-energy input and causes poor stability and portability. To improve the deposited layer quality, conducting online measurements and feedback control of the dimensions, temperature, and other melt-pool parameters during deposition is essential. Currently, melt-pool dimension measurement is mainly based on machine vision methods, which can mostly detect only the deposition direction of a single melt pool, limiting their measurement range and applicability. We propose a binocular-vision-based online measurement method to detect the melt-pool width during DLD. The method uses a perspective transformation algorithm to align multicamera measurements into a single-coordinate system and a fuzzy entropy threshold segmentation algorithm to extract the melt-pool true contour. This effectively captures melt-pool width information in various deposition directions. A DLD measurement system was constructed, establishing an online model that maps the melt-pool width to the offline deposited layer width, validating the accuracy of the binocular vision system in measuring melt-pool width at different deposition angles. The method achieved high accuracy for melt-pool measurements within certain deposition angle ranges. Within the 30°–60° measurement range, the average error is 0.056 mm, with &lt;3% error. The proposed method enhances the detectable range of melt-pool widths, improving cladding layers and parts.

A research on the relationship between landslide area changes and environmental factors in the southern Tibetan plateau

Introduction: Landslides are known to be one of the most frequent types of geological disasters. However, there is not an established method for large-scale, rapid, and high-precision landslide extraction. The quantitative impact of environmental changes on landslide development is also not well understood, which hinders accurate assessments and decision-making in environmental and disaster response. The polar regions, including the Antarctic, the Arctic, and the Tibetan plateau (TP), sensitive to global environmental changes, are significantly affected by global warming. This leads to extensive landslide development, particularly in the southern TP. This research focuses on new landslides in the southern TP, exploring extraction methods and the relationship between landslides and environmental factors.Methods: Utilizing the Google Earth Engine (GEE) and an improved Otsu threshold segmentation algorithm, we processed remote sensing images with 10 m resolution to identify landslide areas. The proposed Normalized Landslide Bare-soil Separation Index (NDLBSI) achieved an 87% pre-extraction accuracy in extracting landslides from Sentinel-2 images from 2019 to 2023. For the pre-extraction results, manual interpretation and correction were carried out, and a model correlating annual landslide changes with environmental factors was established based on least squares multivariate statistical methods.Results: Results show that a significant increase in landslide areas in the southern TP over the past 5 years, correlating with the watershed-wide increase in annual average temperature and vegetation cover, along with a decrease in snow cover area.Discussion: These changes could affect soil and rock moisture, influencing soil stability and landslide occurrence. The study provides valuable insights for large-scale landslide detection and understanding the environmental factors influencing landslides, which is of some significance for landslide hazards early warning.

Estimation of Maize Residue Cover Using Remote Sensing Based on Adaptive Threshold Segmentation and CatBoost Algorithm

Maize residue cover (MRC) is an important parameter to quantify the degree of crop residue cover in the field and its spatial distribution characteristics. It is also a key indicator of conservation tillage. Rapid and accurate estimation of maize residue cover (MRC) and spatial mapping are of great significance to increasing soil organic carbon, reducing wind and water erosion, and maintaining soil and water. Currently, the estimation of maize residue cover in large areas suffers from low modeling accuracy and poor working efficiency. Therefore, how to improve the accuracy and efficiency of maize residue cover estimation has become a research hotspot. In this study, adaptive threshold segmentation (Yen) and the CatBoost algorithm are integrated and fused to construct a residue coverage estimation method based on multispectral remote sensing images. The maize planting areas in and around Sihe Town in Jilin Province, China, were selected as typical experimental regions, and the unmanned aerial vehicle (UAV) was employed to capture maize residue cover images of sample plots within the area. The Yen algorithm was applied to calculate and analyze maize residue cover. The successive projections algorithm (SPA) was used to extract spectral feature indices from Sentinel-2A multispectral images. Subsequently, the CatBoost algorithm was used to construct a maize residue cover estimation model based on spectral feature indices, thereby plotting the spatial distribution map of maize residue cover in the experimental area. The results show that the image segmentation based on the Yen algorithm outperforms traditional segmentation methods, with the highest Dice coefficient reaching 81.71%, effectively improving the accuracy of maize residue cover recognition in sample plots. By combining the spectral index calculation with the SPA algorithm, the spectral features of the images are effectively extracted, and the spectral feature indices such as NDTI and STI are determined. These indices are significantly correlated with maize residue cover. The accuracy of the maize residue cover estimation model built using the CatBoost model surpasses that of traditional machine learning models, with a maximum determination coefficient (R2) of 0.83 in the validation set. The maize residue cover estimation model constructed based on the Yen and CatBoost algorithms effectively enhances the accuracy and reliability of estimating maize residue cover in large areas using multispectral imagery, providing accurate and reliable data support and services for precision agriculture and conservation tillage.

Action capture and VR interactive system for online experimental teaching

The development of technology has intensified people’s interest in virtual reality technology. Virtual reality technology is also widely used in fields such as teaching and healthcare. To improve the real-time and operability of virtual reality interaction systems, a motion capture and virtual reality interaction system design for online experimental teaching is proposed. The active apparent model is used for facial feature point localization, and the joint capture method is improved by combining threshold segmentation and forward kinematics algorithms. Experimental data confirms that compared to Kinect method, the improved hand joint detection method has higher joint positioning accuracy, with a loss rate of less than 40 % for hand joints. The average judgment accuracy of the six facial expression tests is 82 %, 81 %, 79 %, 78 %, 80 %, and 81 %, respectively. The comprehensive recognition accuracy of facial expression recognition is 80.17 %. The read, write, and update times for each frame of the virtual reality interaction system are 102.6 ms and 427 ms, respectively. The system memory usage is 920.5 M, and the CPU and GPU usage rates are 25.2 % and 4.4 %, respectively. The system can run smoothly. The virtual reality system can run continuously for 24 h, with strong operability and low difficulty in expanding system functions, achieving design goals.

Online detection of eggshell cracks using adaptive light‐source intensity

AbstractUnder the fixed light source intensity (LSI), eggshell thickness difference results in excessive exposure or darkening in imaging, which affects the accuracy of crack identification. To optimize it, a solution involves the utilization of an innovative adaptive light source for optical imaging is proposed. A multi function station is adopted. Among them, one station serves as a reference workstation is utilized to obtain the average grayscale value of the eggs image. The other sub stations serve as detection workstations to acquire and detect the panoramic images of the eggshell. By using the red component average grayscale as a reference to adjust the LSI in the remaining sub stations, imaging consistency is effectively. Moreover, a dual threshold segmentation algorithm was developed to attenuate the dark spot, thereby improving crack detection. Experiments results demonstrating the proposed method significantly improving the detection accuracy of non microcracked eggs while achieving approximately 90% and 92% detection accuracy for intact and cracked eggs, respectively.Practical ApplicationsThere are a few studies on the consistency of light for eggs, and in actual online detection, the light source is often set to a fixed light intensity value, which can lead to overexposure or over‐darkening of the egg image due to variations in the eggshell thickness, and both overexposed and over‐darkened images will result in decreased accuracy of crack recognition. This study designed and built an online detection system for eggshell cracks based on variable light intensity. Therefore, the system not only can help to improve the crack recognition rate of poultry eggs and improve their quality but also would help to promote the development of the commercialization of poultry eggs, which has a broad application prospect.

Tool Damage Detection Method Based on Improved Threshold Segmentation Algorithm

In order to solve the problem that the current tool wear defects are difficult to be collected by the visual inspection system, an Otsu threshold segmentation algorithm based on particle swarm optimization is proposed.to detect tool wear The algorithm improved update strategy for inertia coefficients which effectively expanding the search scope of the algorithm and shortens the running time of the algorithm. By adding a perturbation equation to the particle swarm, solved the problem of traditional particle swarm optimization algorithms easily falling into local optima. Finally, an experimental platform is built to verify the effectiveness of the algorithm. This inspection method can achieve the identification of tool damage areas and the measurement of tool damage amount, and has advantages such as high recognition accuracy and fast running speed compared to traditional Otsu algorithm, Canny algorithm ,local threshold segmentation and so on. The research results have certain reference value for the actual tool defect detection system.

Quantitatively characterizing sandy soil structure altered by MICP using multi-level thresholding segmentation algorithm

The influences of biological, chemical, and flow processes on soil structure through microbial-induced carbonate precipitation (MICP) are not yet fully understood. In this study, we use a multi-level thresholding segmentation algorithm, genetic algorithm (GA) enhanced Kapur entropy (KE) (GAE-KE), to accomplish quantitative characterization of sandy soil structure altered by MICP cementation. A sandy soil sample was treated using MICP method and scanned by the synchrotron radiation (SR) micro-CT with a resolution of 6.5 μm. After validation, tri-level thresholding segmentation using GAE-KE successfully separated the precipitated calcium carbonate crystals from sand particles and pores. The spatial distributions of porosity, pore structure parameters, and flow characteristics were calculated for quantitative characterization. The results offer pore-scale insights into the MICP treatment effect, and the quantitative understanding confirms the feasibility of the GAE-KE multi-level thresholding segmentation algorithm.

Threshold segmentation based on information fusion for object shadow detection in remote sensing images

Threshold segmentation based on information fusion for object shadow detection in remote sensing images

Multilayer grid XG Boost architecture based automatic osteosarcoma classification

The application of machinelearning(ML) and deep learning (DL)methods might reduce the amount of time required by clinicians while simultaneously enhancing patient outcomes. The classification algorithm has to be provided with a huge amount of data in order to improve its accuracy. In this study, a combination of ML and DL is utilized to differentiate between images of normal and necrotic tissues utilizing a public database of osteosarcoma histological images. The data was initially preprocessed, and contour based threshold segmentation algorithms were performed. Next, the anomalous features are extracted using stochastic linear embedding-based feature extraction. Finally, stained photos are used to train the proposed multilayer grid XG Boost classifier, which improves output accuracy. The results of the experiments show that the suggested classifier is the most accurate one presently in use for categorizing illnesses. With H and E stained images, our improved model performed better at detecting osteosarcoma malignancy.

Detection of early bruises on apples using near‐infrared camera imaging technology combined with adaptive threshold segmentation algorithm

AbstractApples are highly susceptible to bruising during picking, packing, and transportation, and early detection of apple bruises allows timely screening and treatment to reduce food safety risks. However, early bruises on apples, especially bruises of 30 min or less, with similar appearance to sound tissue, making manual sorting difficult. To address this problem, a near‐infrared (NIR) camera imaging technology combined with adaptive threshold segmentation algorithm for early apple bruise detection is proposed in this article. First, a NIR camera imaging system is built to acquire full‐band images of fruit face, fruit stem face and calyx face of sound apples and apples with single bruise and multiple bruises at 30 min of occurrence. Then, an adaptive threshold segmentation algorithm is proposed and an image processing algorithm, which includes high‐pass filtering, feature extraction, adaptive threshold segmentation and Hough circle detection, is developed and applied to full‐band images in response to the interference of apple stem and calyx on early bruising detection in this article. The results show that the proposed algorithm has an accuracy of 95.56% for the detection of sound apple samples and an F1‐score of 94.70% for the detection of early bruised samples, which is better than the commonly used image segmentation algorithms such as Otsu and fixed thresholding. This study shows that the combination of NIR camera imaging and the adaptive threshold segmentation algorithm proposed in this article can effectively detect early bruises on apples, providing a new idea for rapid and nondestructive detection of agricultural product quality.Practical applicationsDetecting early bruises on apples can achieve effective monitoring and control of apple supply chain, and guarantee the quality and safety of agricultural products. Most studies are based on conventional machine vision with RGB camera or hyperspectral imaging technology, while conventional machine vision based on RGB camera is difficult to detect apple early bruises that are not visible to the naked eye, and hyperspectral imaging has a large amount of data and redundant image data among wavelengths. Compared with previous studies, apple early bruise detection method proposed in this article combining NIR camera imaging technology and adaptive threshold segmentation algorithm can avoid the interference of bright spots, fruit stem and calyx on bruise segmentation while capturing a single channel image, providing a new method for accurate and rapid detection of early bruises on apples.

Comparison of Multiple Segmentation Methods for Volumetric Delineation of Primary Prostate Cancer with Prostate-Specific Membrane Antigen-Targeted 18F-DCFPyL PET/CT.

This study aimed to assess the accuracy of intraprostatic tumor volume measurements on prostate-specific membrane antigen-targeted 18F-DCFPyL PET/CT made with various segmentation methods. An accurate understanding of tumor volumes versus segmentation techniques is critical for therapy planning, such as radiation dose volume determination and response assessment. Methods: Twenty-five men with clinically localized, high-risk prostate cancer were imaged with 18F-DCFPyL PET/CT before radical prostatectomy. The tumor volumes and tumor-to-prostate ratios (TPRs) of dominant intraprostatic foci of uptake were determined using semiautomatic segmentation (applying SUVmax percentage [SUV%] thresholds of SUV30%-SUV70%), adaptive segmentation (using adaptive segmentation percentage [A%] thresholds of A30%-A70%), and manual contouring. The histopathologic tumor volume (TV-Histo) served as the reference standard. The significance of differences between TV-Histo and PET-based tumor volume were assessed using the paired-sample Wilcoxon signed-rank test. The Spearman correlation coefficient was used to establish the strength of the association between TV-Histo and PET-derived tumor volume. Results: Median TV-Histo was 2.03 cm3 (interquartile ratio [IQR], 1.16-3.36 cm3), and median TPR was 10.16%. The adaptive method with an A40% threshold most closely determined the tumor volume, with a median difference of +0.19 (IQR, -0.71 to +2.01) and a median relative difference of +7.6%. The paired-sample Wilcoxon test showed no significant difference in PET-derived tumor volume and TV-Histo using A40%, A50%, SUV40%, and SUV50% threshold segmentation algorithms (P > 0.05). For both threshold-based segmentation methods, use of higher thresholds (e.g., SUV60% or SUV70% and A50%-A70%) resulted in underestimation of tumor volumes, and use of lower thresholds (e.g., SUV30% or SUV40% and A30%) resulted in overestimation of tumor volumes relative to TV-Histo and TPR. Manual segmentation overestimated the tumor volume, with a median difference of +2.49 (IQR, 0.42-4.11) and a median relative difference of +130%. Conclusion: Segmentation of intraprostatic tumor volume and TPR with an adaptive segmentation approach most closely approximates TV-Histo. This information might be used to guide the primary treatment of men with clinically localized, high-risk prostate cancer.

A novel method for seed cotton color measurement based on machine vision technology

Color of seed cotton is one of the key indexes of seed cotton quality, which greatly affects the price, grading, storage, and processing of seed cotton. Currently, there are shortage of mature color measurement methods and equipment specifically for seed cotton. Therefore, a color measurement method for seed cotton based on machine vision technology was proposed in this research. To solve the problem of color difference in images, a color difference correction algorithm based on multiple linear regression was proposed, which significantly reduced the color difference by 54.19%. To segment large impurities and large hard particles (cotton seeds, cotton stalks, and boll shells) that are easy to produce shadows from seed cotton images, a quadratic dynamic thresholding segmentation algorithm based on multi-channel fusion was proposed, which significantly improved the segmentation accuracy. The verification results showed that the average value of the intersection over union was 0.9. In the calculation of the color indexes of seed cotton, a correction algorithm based on the BP neural network was used to correct the indexes by taking standard tiles as a reference to reduce the difference caused by system error. The results of the machine vision method were compared with those of the detection of corresponding lint by HVI 1000 and spectrophotometer HX-410. The coefficients of determination (R2) of the Reflectance degree (Rd) and Yellowness (+b) measured by HVI 1000 were 0.790 and 0.865, respectively. The R2 for Rd and +b measured by HX-410 were 0.809 and 0.879, respectively. In addition, the analysis results of the effect of impurities and shadows on seed cotton color showed that both impurities and shadows had a negative effect on Rd. However, the effect of shadows on +b was negative and the effect of impurities was positive. This study indicated that it was feasible to detect seed cotton color using machine vision method.

Research on Obstacle Detection and Avoidance of Autonomous Underwater Vehicle Based on Forward-Looking Sonar.

Due to the complexity of the ocean environment, an autonomous underwater vehicle (AUV) is disturbed by obstacles when performing tasks. Therefore, the research on underwater obstacle detection and avoidance is particularly important. Based on the images collected by a forward-looking sonar on an AUV, this article proposes an obstacle detection and avoidance algorithm. First, a deep learning-based obstacle candidate area detection algorithm is developed. This algorithm uses the You Only Look Once (YOLO) v3 network to determine obstacle candidate areas in a sonar image. Then, in the determined obstacle candidate areas, the obstacle detection algorithm based on the improved threshold segmentation algorithm is used to detect obstacles accurately. Finally, using the obstacle detection results obtained from the sonar images, an obstacle avoidance algorithm based on deep reinforcement learning (DRL) is developed to plan a reasonable obstacle avoidance path of an AUV. Experimental results show that the proposed algorithms improve obstacle detection accuracy and processing speed of sonar images. At the same time, the proposed algorithms ensure AUV navigation safety in a complex obstacle environment.

Multilevel Thresholding Segmentation based on Levy-Horse Optimization Algorithm: A Review

Abstract: A novel multi-level thresholding segmentation algorithm is proposed concerning machine learning and a metaheuristic algorithm is used to select multiple threshold value for segmenting the image. Here for the analysis, two, three and five-level of threshold value for the segmentation process is made. The threshold values for two, three and five level is generated by Levy horse optimized support vector machine (LHSVM). This optimized machine learning approach selects a reduced error threshold value to segment every object or region of interest in an image. Then the method is evaluated using Berkeley segmentation dataset and benchmark (BSDS300) dataset. This kind of multi-level threshold segmentation splits the image into class and thereby considers several classes for segmenting each object in the image

A Method for Extracting Contours of Building Facade Hollowing Defects Using Polarization Thermal Images Based on Improved Canny Algorithm

During the service process of high-rise buildings, hollowing defects may be produced in the decorative layer, which not only affect the appearance, but also create a safety hazard of wall covering and shattered plaster peeling. Numerous studies have shown that hollowing can be detected using infrared thermal imagery under normal conditions. However, it is difficult to detect the edge and calculate the area of the hollowing on an exterior facade accurately because of the low contrast and fuzzy boundaries of the obtained infrared thermal images. To address these problems, a method for extracting the contours of building facade hollowing defects using polarization thermal images based on an improved Canny algorithm has been proposed in this paper. Firstly, the principle of thermal polarization imaging was introduced for hollowing detection. Secondly, considering the shortcomings of the Canny edge detection algorithm and the features of polarization thermal images, an improved Canny edge detection algorithm is proposed, including adaptive bilateral filtering to improve noise reduction ability while ensuring defect edges are not virtualized, Laplacian sharpening and histogram equalization to achieve contour sharpening and contrast enhancement, and eight-direction gradient templates for calculating image gradients, which make interpolation with non-maximum suppression more accurate, and the Tsallis entropy threshold segmentation algorithm based on the OTSU algorithm verification makes the image contour information more complete and accurate. Finally, a long-wave infrared polarization thermal imaging experimental platform was established and validation experiments were conducted. The experimental results demonstrate that the distinct, smooth, and precise location edges of the hollowing polarization infrared thermal images can be obtained, and the average error of the detected hollowing area is about 10% using the algorithm proposed in this paper.

Molten pool image processing and quality monitoring of laser cladding process based on coaxial vision

Aiming at the problems of the traditional detection method of cladding quality in laser cladding process, which relies on experience, has low accuracy and long recognition time, a method based on the image area of molten pool is proposed to detect cladding quality. Firstly, the median filter algorithm with dynamic adjustment of filter window is used to reduce noise and the iterative threshold segmentation algorithm is used to separate the molten pool from the boundary, then the edge of the molten pool is extracted by Canny algorithm which can adaptively change the high and low thresholds and finally the area value of the molten pool is obtained according to the correspondence between the pixel value and the actual value. The relationship between three key factors (nozzle angle, scanning speed and laser power) and molten pool area is analyzed to achieve favorable cladding quality. The results show that the influence of nozzle angle on the molten pool area is anisotropic. And the molten pool area decreases monotonically with the increase of scanning speed, while the higher velocity will destroy the formation of the molten pool. In addition, with the increase of laser power, the expansion of molten pool area is intensified. At the same time, the analysis of the change of molten pool area over time can quickly and accurately detect the three defects (protrusion, depression, and slag inclusion) in the process of cladding.

Drought recognition based on feature extraction of multispectral images for the soybean canopy

Drought stress is a major environmental factor in soybean growth. Accurate and rapid detection of phenotypic traits during drought condition is important for water-saving irrigation and variety selection for the drought-resistant plant. The traditional manual measurement of phenotypes is time-consuming and error-prone, thus, an accurate and rapid drought recognition method was proposed based on feature extraction of multispectral images for the soybean canopy. First, the Suinong26 soybean variety was selected as the research object. The multispectral sensor (Parrot Drone SAS, Paris, France) was used to acquire the soybean canopies' multispectral image in four channels (NIR, RED, REG, and GRE). Second, after de-noising the original image using median filtering method, the target image of the soybean canopy in the NIR channel was extracted using a thresholding segmentation algorithm. Third, the segmented image was set as an image template to perform canopy segmentation on the rest of the channel images, the soybean canopy images were obtained in the RED, REG, and GRE channels. In addition, soybean canopies' 9 vegetation indexes (NDVI, GNDVI, NDGI, RVI, DVI, GCI, RECI, RDVI, NLI) and each channel's 7 texture features (mean, standard deviation, smoothness, third moment, information entropy, average gradient, fractal dimension) were calculated by using image fusion and processing technologies. Finally, based on the 37-dimensional features of the extracted multispectral images of the soybean canopy, a drought recognition model was established using a support vector machine (SVM) with a radial basis as the kernel function. The models' recognition accuracy reached 96.87%, and the running time was only 0.493441 s. The proposed model has improved the recognition accuracy by 8.33% on average compared with Genetic Algorithm-Backpropagation (GA-BP) neural network, radial basis neural network (RBF) and random forest (RF) method. The proposed method can provide a theoretical basis and technical support for precise regulation and scientific decision-making of breeding and cultivation management for drought-resistant soybean varieties.

Magnesium Ingot Stacking Segmentation Algorithm for Industrial Robot Based on the Correction of Image Overexposure Area.

This paper proposes an adaptive threshold segmentation algorithm for the magnesium ingot stack based on image overexposure area correction (ATSIOAC), which solves the problem of mirror reflection on the surface of magnesium alloy ingots caused by external ambient light and auxiliary light sources. Firstly, considering the brightness and chromaticity information of the mapped image, we divide the exposure probability threshold into weak exposure and strong exposure regions. Secondly, the saturation difference between the magnesium ingot region and the background region is used to obtain a mask for the magnesium ingot region to eliminate interference from the image background. Then, the RGB average of adjacent pixels in the overexposed area is used as a reference to correct the colors of the strongly exposed and weakly exposed areas, respectively. Furthermore, in order to smoothly fuse the two corrected images, pixel weighted average (WA) is applied. Finally, the magnesium ingot sorting experimental device was constructed and the corrected top surface image of the ingot pile was segmented through ATSIOAC. The experimental results show that the overexposed area detection and correction algorithm proposed in this paper can effectively correct the color information in the overexposed area, and when segmenting ingot images, complete segmentation results of the top surface of the ingot pile can be obtained, effectively improving the accuracy of magnesium alloy ingot segmentation. The segmentation algorithm achieves a segmentation accuracy of 94.38%.