Robust Clustering Algorithm Research Articles

Scalable Clustering: Large Scale Unsupervised Learning of Gaussian Mixture Models with Outliers

Clustering is a widely used technique with a long and rich history in a variety of areas. However, most existing algorithms do not scale well to large datasets, or are missing theoretical guarantees of convergence. This paper introduces a provably robust clustering algorithm based on loss minimization that performs well on Gaussian mixture models with outliers. It provides theoretical guarantees that the algorithm obtains high accuracy with high probability under certain assumptions. Moreover, it can also be used as an initialization strategy for k-means clustering. Experiments on real-world large-scale datasets demonstrate the effectiveness of the algorithm when clustering a large number of clusters, and a k-means algorithm initialized by the algorithm outperforms many of the classic clustering methods in both speed and accuracy, while scaling well to large datasets such as ImageNet.

Plasma Cytokines Pattern as a Prognostic Marker for Esophageal Squamous Cell Carcinoma via Unsupervised Clustering Analyses

Introduction: Cytokines such as tumor necrosis factor-alpha (TNF-α), interleukin 6 (IL6), interferon-gamma (IFN-γ), interleukin 17-alpha (IL17-α), and interleukin 33 (IL33) play critical roles in immune responses and may impact cancer prognosis in future. However, few studies have simultaneously explored the prognostic impact of these cytokines for cancer. In this study, we aim to apply the unsupervised clustering analysis to approach the correlation between the expression of these cytokines and the subsequent prognosis of patients with esophageal squamous cell carcinoma (ESCC). Methods: A robust clustering algorithm was used, the Gaussian mixture method (GMM), through the mclust R package to group patients based on the expression of their cytokines in plasma or tumors. The 324 NTU patients were grouped into 4 clusters, and the 179 GSE53625 patients were grouped into 3 clusters based on expression in plasma and tumors, respectively. Five- and 3-year overall survival (OS) and progression-free survival (PFS) curves of each cluster were compared. Univariate and multivariate Cox regression analyses were also performed. Results: We successfully distinguished the multimodal distribution of cytokines through GMM clustering and discovered the relationship between cytokines and clinical outcomes. We observed that NTU-G3 and NTU-G4 subgroups showed most variation in 5-, 3-year OS and 5-, 3-year PFS with NTU-G3 being associated with poorer prognosis compared to NTU-G4 (p = 0.016, 0.0052, 0.0575, and 0.0168, respectively). NTU-G3 was characterized with higher TNF-α (median = 3.855, N = 78) and lower IL33 (median = 0.000, N = 78), while NTU-G4 showed lower TNF-α (median = 1.76, N = 51) and higher IL33 (median = 1.070, N = 51). The difference was statistically significant for TNF-α and IL33, with p = 0.0002 and p < 0.0001, respectively. A multivariate Cox-regression analysis revealed that GMM clustering and T/N stage were independent factors for prognosis, suggesting that the prognosis might be dependent on these cytokines. Conclusions: Our data suggest that expression patterns of IL33 and TNF-α in plasma might serve as a convenient marker to predict the prognosis of ESCC in the future.

Benchmarking clustering, alignment, and integration methods for spatial transcriptomics

BackgroundSpatial transcriptomics (ST) is advancing our understanding of complex tissues and organisms. However, building a robust clustering algorithm to define spatially coherent regions in a single tissue slice and aligning or integrating multiple tissue slices originating from diverse sources for essential downstream analyses remains challenging. Numerous clustering, alignment, and integration methods have been specifically designed for ST data by leveraging its spatial information. The absence of comprehensive benchmark studies complicates the selection of methods and future method development.ResultsIn this study, we systematically benchmark a variety of state-of-the-art algorithms with a wide range of real and simulated datasets of varying sizes, technologies, species, and complexity. We analyze the strengths and weaknesses of each method using diverse quantitative and qualitative metrics and analyses, including eight metrics for spatial clustering accuracy and contiguity, uniform manifold approximation and projection visualization, layer-wise and spot-to-spot alignment accuracy, and 3D reconstruction, which are designed to assess method performance as well as data quality. The code used for evaluation is available on our GitHub. Additionally, we provide online notebook tutorials and documentation to facilitate the reproduction of all benchmarking results and to support the study of new methods and new datasets.ConclusionsOur analyses lead to comprehensive recommendations that cover multiple aspects, helping users to select optimal tools for their specific needs and guide future method development.

Robust clustering algorithm: The use of soft trimming approach

The presence of noise or outliers in data sets may heavily affect the performance of clustering algorithms and lead to unsatisfactory results. The majority of conventional clustering algorithms are sensitive to noise and outliers. Robust clustering algorithms often overcome difficulties associated with noise and outliers and find true cluster structures. We introduce a soft trimming approach for the hard clustering problem where its objective is modeled as a sum of the cluster function and a function represented as a composition of the algebraic and distance functions. We utilize the composite function to estimate the degree of the significance of each data point in clustering. A robust clustering algorithm based on the new model and a procedure for generating starting cluster centers is developed. We demonstrate the performance of the proposed algorithm using some synthetic and real-world data sets containing noise and outliers. We also compare its performance with that of some well-known clustering techniques. Results show that the new algorithm is robust to noise and outliers and finds true cluster structures.

Accelerated Stochastic Variance Reduction Gradient Algorithms for Robust Subspace Clustering.

Robust face clustering enjoys a wide range of applications for gate passes, surveillance systems and security analysis in embedded sensors. Nevertheless, existing algorithms have limitations in finding accurate clusters when data contain noise (e.g., occluded face clustering and recognition). It is known that in subspace clustering, the ℓ1- and ℓ2-norm regularizers can improve subspace preservation and connectivity, respectively, and the elastic net regularizer (i.e., the mixture of the ℓ1- and ℓ2-norms) provides a balance between the two properties. However, existing deterministic methods have high per iteration computational complexities, making them inapplicable to large-scale problems. To address this issue, this paper proposes the first accelerated stochastic variance reduction gradient (RASVRG) algorithm for robust subspace clustering. We also introduce a new momentum acceleration technique for the RASVRG algorithm. As a result of the involvement of this momentum, the RASVRG algorithm achieves both the best oracle complexity and the fastest convergence rate, and it reaches higher efficiency in practice for both strongly convex and not strongly convex models. Various experimental results show that the RASVRG algorithm outperformed existing state-of-the-art methods with elastic net and ℓ1-norm regularizers in terms of accuracy in most cases. As demonstrated on real-world face datasets with different manually added levels of pixel corruption and occlusion situations, the RASVRG algorithm achieved much better performance in terms of accuracy and robustness.

Robust Fuzzy C-Means Clustering Algorithm Based on Normal Shrink and Membership Filtering for Image Segmentation

The robustness and effectiveness of image segmentation using the FCM algorithm can be improved by incorporating local spatial information into the FCM method, which is particularly noise-tolerant. However the introduction of local spatial information gives more computational complexity. Hence to overcome this problem an improved FCM clustering method is proposed which is based on a normal shrink algorithm with membership filtering. The Proposed method gives a faster and more robust result in comparison to FCM. Firstly, a normal shrink denoising algorithm is introduced to preserve the image details and noise immunity. Secondly, membership filtering is introduced, which depends only on the local spatial neighboring properties of the matrix called the membership partition matrix. The Proposed method is faster and simpler as it does not calculate the distance between pixels and cluster centers and between local spatial neighboring. Also, it is very efficient for noise immunity.

Joint learning framework of superpixel generation and fuzzy sparse subspace clustering for color image segmentation

Sparse subspace clustering (SSC) is an important image segmentation method that constructs a self-representation coefficient matrix to represent the relationships between pixels, and then uses spectral clustering to achieve clustering. Compared with other unsupervised segmentation algorithms, SSC has good segmentation performance. However, SSC has a high computational complexity when processing large-sized image. To improve computational efficiency, many researchers have proposed superpixel-based SSC algorithms, which process superpixels instead of pixels and improve efficiency through preprocessing. Due to the sensitivity of superpixel generation to noise, superpixel-based SSC algorithms still have poor robustness. Additionally, preprocessing increases the complexity of the algorithm. To address these issues, this paper proposes a robust superpixel-based fuzzy sparse subspace clustering algorithm. This algorithm combines fuzzy sparse subspace clustering with superpixel generation, and it constructs a unified optimization learning framework through fuzzy C-multiple-means clustering to improve segmentation performance and reduce complexity. Additionally, this paper introduces additional features of superpixel in sparse subspace clustering to further enhance the segmentation performance of the algorithm. Experiment results indicate that the proposed algorithm not only outperforms existing state-of-the-art robust segmentation algorithms independent of superpixels, but also is superior to the latest superpixel-based segmentation algorithms.

Application of machine learning for identification of heterotic groups in sunflower through combined approach of phenotyping, genotyping and protein profiling

Application of machine learning in plant breeding is a recent concept, that has to be optimized for precise utilization in the breeding program of high yielding crop plants. Identification and efficient utilization of heterotic grouping pattern aided with machine learning approaches is of utmost importance in hybrid cultivar breeding as it can save time and resources required to breed a new plant hybrid/variety. In the present study, 109 genotypes of sunflower were investigated at morphological, biochemical (SDS-PAGE) and molecular levels (through micro-satellites (SSR) markers) for heterotic grouping. All the three datasets were combined, scaled, and subjected to unsupervised machine learning algorithms, i.e., Hierarchical clustering, K-means clustering and hybrid clustering algorithm (hierarchical + K-means) for assessment of efficiency and resolution power of these algorithms in practical plant breeding for heterotic grouping identification. Following the application of machine learning unsupervised clustering approach, two major groups were identified in the studied sunflower germplasm, and further classification revealed six smaller classes in each major group through hierarchical and hybrid clustering approach. Due to high resolution, obtained in hierarchical clustering, classification achieved through this algorithm was further used for selection of potential parents. One genotype from each smaller group was selected based on the maximum seed yield potential and hybridized in a line × tester mating design producing 36 F1 cross combinations. These F1s along with their parents were studied in open field conditions for validating the efficacy of identified heterotic groups in sunflowers genetic material under study. Data for 11 agronomic and qualitative traits were recorded. These 36 F1 combinations were tested for their combining ability (General/Specific), heterosis, genotypic and phenotypic correlation and path analysis. Results suggested that F1 hybrids performed better for all the traits under investigation than their respective parents. Findings of the study validated the use of machine learning approaches in practical plant breeding; however, more accurate and robust clustering algorithms need to be developed to handle the data noisiness of open field experiments.

Fast and robust clustering of general-shaped structures with tk-merge

In real-world applications, the group of provenance of data can be inherently uncertain, the data values can be imprecise and some of them can be wrong. We handle uncertain, imprecise and noisy data in clustering problems with general-shaped structures. We do it under very weak parametric assumptions with a two-step hybrid robust clustering algorithm based on trimmed k-means and hierarchical agglomeration. The algorithm has low computational complexity and effectively identifies the clusters also in presence of data contamination. We also present natural generalizations of the approach as well as an adaptive procedure to estimate the amount of contamination in a data-driven fashion. Our proposal outperforms state-of-the-art robust, model-based methods in our numerical simulations and real-world applications related to color quantization for image analysis, human mobility patterns based on GPS data, biomedical images of diabetic retinopathy, and weather data.

A neighborhood-based robust clustering algorithm using Apollonius function kernel

Density Peak Clustering (DPC) is a useful algorithm for grouping data points based on their closeness and density. It is widely used in various areas of computer science. It can detect the peak density in each period of the clustering steps and assign the rest of the data to the next steps. Peak density clustering analysis is able to discover clusters of arbitrary shapes, but usually, when the data set is large, it is computationally expensive and difficult to apply to high-dimensional data. When the distance between clusters is significantly different, it is unable to distinguish adjacent clusters with different densities. To deal with such issues, a new clustering algorithm based on neighborhoods and the Apollonius kernel is proposed. The proposed algorithm can be implemented with four steps. 1) Estimation of natural neighbor, 2) Identification of core points, 3) Merging the cores of a cluster and separating the cores of different clusters, 4) Using the Apollonius kernel between the core points and the farthest points to assign the remaining points to the clusters and find the noise points. The local density estimation process based on the natural neighborhood of high-density points and their integration, as well as border point classification, allows the proposed method to reveal the local cluster structure. As a result, the adjoining clusters with distinctive densities are recognized, and the computation is substantially reduced. The proposed method is validated on some synthetic data sets and the UCI data set experimentally and has a high 10% accuracy.

Multi-view [formula omitted]-means clustering algorithm based on redundant and sparse feature learning

In recent years, multi-view clustering has been widely concerned and studied. Being its data generally has high dimension and noise, how to effectively deal with redundant features in data from various views and improve the clustering effect is an important issue in multi-view clustering. In this paper, a robust multi-view clustering algorithm Multi-view K-means clustering algorithm based on redundant and sparse feature learning (RSFMVKM) considering redundancy and sparse feature learning is proposed, which can effectively reduce the dimensions of data from different views. Firstly, the K-means algorithm is extended to multi-view clustering, and the importance of each views is represented by learning the weight of each view. Secondly, the sparse feature representation is obtained by applying the norm constraint to the projection matrix in a single view, and the cosine similarity is used to depict the redundancy matrix, which can remove the redundant information and improve the clustering effect. Moreover, the proposed algorithm can effectively avoid the generation of empty classes, and the clustering index matrix is discrete, so the clustering results can be obtained directly. Finally, the proposed algorithm is compared with six state-of-the-art multi-view clustering algorithms on six public datasets, and the experimental results show that the algorithm is effective.

Optimizing environmental monitoring in IoT: integrating DBSCAN with genetic algorithms for enhanced clustering

In our study, we introduce an advanced clustering method designed for IoT-based environmental monitoring. We’ve combined two powerful techniques, Density-Based Spatial Clustering of Applications with Noise (DBSCAN) and Genetic Algorithms (GA), to create a specialized approach called EC-GAD (Enhanced-Clustering using Genetic Algorithms and DBSCAN). This integrated system model relies on DBSCAN, a robust clustering algorithm capable of handling irregular shapes and varying data densities, to group sensor nodes based on their physical proximity. To improve clustering performance, we’ve harnessed Genetic Algorithms to optimize the parameters of DBSCAN. Through a repetitive process involving selection, crossover, and mutation, GA refines parameter settings based on the quality of environmental clustering as assessed by fitness metrics. Our approach is tailored specifically for IoT deployments in environmental monitoring, which involve collecting data from sensor nodes and integrating DBSCAN and GA. We’ve paid special attention to choosing an appropriate distance metric and fine-tuning DBSCAN parameters such as epsilon (ε) and minPts to match the unique needs of environmental monitoring applications. Furthermore, we’ve taken energy efficiency into account by implementing energy-aware node selection and optimizing cluster formation to minimize energy consumption.

Multiomics characterization and verification of clear cell renal cell carcinoma molecular subtypes to guide precise chemotherapy and immunotherapy.

Clear cell renal cell carcinoma (ccRCC) is a heterogeneous tumor with different genetic and molecular alterations. Schemes for ccRCC classification system based on multiomics are urgent, to promote further biological insights. Two hundred and fifty-fiveccRCC patients with paired data of clinical information, transcriptome expression profiles, copy number alterations, DNA methylation, and somatic mutations were collected for identification. Bioinformatic analyses were performed based on our team's recently developed R package "MOVICS."With 10 state-of-the-art algorithms, we identified the multiomics subtypes (MoSs) for ccRCC patients. MoS1 is an immune exhausted subtype, presented the poorest prognosis, and might be caused by an exhausted immune microenvironment, activated hypoxia features, but can benefit from PI3K/AKT inhibitors. MoS2 is an immune "cold" subtype, which represented more mutation of VHL and PBRM1, favorable prognosis, and is more suitable for sunitinib therapy. MoS3 is the immune "hot" subtype, and can benefit from the anti-PD-1 immunotherapy. We successfully verified the different molecular features of the three MoSs in external cohorts GSE22541, GSE40435, and GSE53573. Patients that received Nivolumab therapy helped us to confirm that MoS3 is suitable for anti-PD-1 therapy. E-MTAB-3267 cohort also supported the fact that MoS2 patients can respond more to sunitinib treatment. We also confirm that SETD2 is a tumor suppressor in ccRCC, along with the decreased SETD2 protein level in advanced tumor stage, and knock-down ofSETD2 leads to the promotion of cell proliferation, migration, and invasion. In summary, we provide novel insights intoccRCC molecular subtypes based on robust clustering algorithms via multiomics data, and encourage future precise treatment ofccRCC patients.

Real-time electron clustering in an event-driven hybrid pixel detector

Event-driven hybrid pixel detectors with nanosecond time resolution have opened up novel pathways in modern ultrafast electron microscopy, for example in hyperspectral electron-energy loss spectroscopy or free-electron quantum optics. However, the impinging electrons typically excite more than one pixel of the device, and an efficient algorithm is therefore needed to convert the measured pixel hits to real single-electron events. Here we present a robust clustering algorithm that is fast enough to find clusters in a continuous stream of raw data in real time. Each tuple of position and arrival time from the detector is continuously compared to a buffer of previous hits until the probability of a merger with an old event becomes irrelevant. In this way, the computation time becomes independent of the density of electron arrival and the algorithm does not break the operation chain. We showcase the performance of the algorithm with a ‘timepix’ camera in two regimes of electron microscopy, in continuous beam emission and laser-triggered femtosecond mode.

GMM-Based Expanded Feature Space as a Way to Extract Useful Information for Rare Cell Subtypes Identification in Single-Cell Mass Cytometry.

Cell subtype identification from mass cytometry data presents a persisting challenge, particularly when dealing with millions of cells. Current solutions are consistently under development, however, their accuracy and sensitivity remain limited, particularly in rare cell-type detection due to frequent downsampling. Additionally, they often lack the capability to analyze large data sets. To overcome these limitations, a new method was suggested to define an extended feature space. When combined with the robust clustering algorithm for big data, it results in more efficient cell clustering. Each marker's intensity distribution is presented as a mixture of normal distributions (Gaussian Mixture Model, GMM), and the expanded space is created by spanning over all obtained GMM components. The projection of the initial flow cytometry marker domain into the expanded space employs GMM-based membership functions. An evaluation conducted on three established cellular identification algorithms (FlowSOM, ClusterX, and PARC) utilizing the most substantial publicly available annotated dataset by Samusik et al. demonstrated the superior performance of the suggested approach in comparison to the standard. Although our approach identified 20 cell clusters instead of the expected 24, their intra-cluster homogeneity and inter-cluster differences were superior to the 24-cluster FlowSOM-based solution.

Robust hesitant fuzzy partitional clustering algorithms and their applications in decision making

Hesitant fuzzy sets (HFSs) are a powerful tool to describe uncertain and vague information, whose relationship can be analyzed and mined by clustering algorithms. The partitional idea is widely used in clustering analysis for real-number data but little for hesitant fuzzy data. After pointing out the weakness of the existing partitional clustering algorithms between HFSs, we modify the existing partitional clustering algorithms and propose four new algorithms with two types under hesitant fuzzy environment. One type, including the hesitant fuzzy K-medians (HFKME) and the hesitant fuzzy C-medians (HFCME) algorithms, is based on median but not mean. The other type, including the improved hesitant fuzzy K-means (IHFKM) and the improved hesitant fuzzy C-means (IFHCM) algorithms, is on the basis of a new and more robust distance measure which will be developed by HFSs. We find that these four new clustering algorithms are more robust than the existing ones and have better performance under the noise environment. After that, we compare our proposed partitional clustering algorithms with each other and also compare them with the existing methods, and then apply them to multi-attribute decision making and cluster analysis. Numerical results show that HFCME has better performance than HFKME, and IHFCM is better than IHFKM.

A Sparse Framework for Robust Possibilistic K-Subspace Clustering

Clustering noisy, high-dimensional, and structurally complex data have always been a challenging task. As most existing clustering methods are not able to deal with both the adverse impact of noisy samples and the complex structures of data, in this article, we propose a novel robust and sparse possibilistic K-subspace (RSPKS) clustering algorithm to integrate subspace recovery and possibilistic clustering algorithms under a unified sparse framework. First, the proposed method sparsifies the membership matrix and the subspace projection vector under a dual-sparse framework to handle high-dimensional noisy data. This unifies dimensionality reduction and clustering using one objective function for which the optimization can be realized through synchronous iteration. Second, the reconstruction error of each sample in the local subspace is used as the distance metric for classification. That is, each sample itself is treated as a clustering prototype so as not to be affected by the structure of the overall data distribution. Therefore, the clustering prototype construction problem of the data with complex structures can be better addressed. Finally, to deal with nonlinear regions, our RSPKS method is further extended into a kernelized version, namely the kernelized RSPKS clustering algorithm. The experimental results on both synthetic and real-world datasets demonstrate that our proposed method outperforms state-of-the-art algorithms in terms of clustering accuracy.

A knee point driven Kriging-assisted multi-objective robust fuzzy clustering algorithm for image segmentation

To enhance the segmentation performance and optimization efficiency of multi-objective evolutionary clustering algorithms for noisy images, this study presents a knee point driven Kriging-assisted multi-objective robust fuzzy clustering algorithm for image segmentation (K2 MORFC). First, pixel and region-level fitness functions are developed to comprehensively consider the spatial constraints and region consistency in an image. In the pixel-level fitness function, a Kullback–Leibler (KL) divergence-based spatial constraint term is employed and combined with intra-class compactness to preserve the membership similarity between an arbitrary image pixel and its non-local neighbors. In the region-level fitness function, image edge information is utilized to evaluate the regional consistency of the partition. Moreover, an adaptive determination mechanism of the weight factor for the spatial constraint term is designed in the pixel-level fitness function to control the influence of the spatial constraint for each pixel. Furthermore, considering that the preference toward knee points can effectively improve the convergence performance, a knee point driven Kriging-assisted evolutionary algorithm (K2EA) is proposed to efficiently optimize these two fitness functions, including a dynamic subspace knee point searching strategy, a knee point guided environmental selection strategy, and a Kriging model management mechanism. Finally, we construct a fuzzy clustering validity index with a KL divergence-based non-local spatial constraint term to select the optimal solution from the nondominated solution set. Competitive experimental results on the DTLZ benchmark verify the effectiveness of K2EA. Segmentation results on color images from the Berkeley dataset and magnetic resonance images from the BrainWeb and IBSR datasets confirm the performance of K2MORFC.

A Robust and High-Dimensional Clustering Algorithm Based on Feature Weight and Entropy.

Since the Fuzzy C-Means algorithm is incapable of considering the influence of different features and exponential constraints on high-dimensional and complex data, a fuzzy clustering algorithm based on non-Euclidean distance combining feature weights and entropy weights is proposed. The proposed algorithm is based on the Fuzzy C-Means soft clustering algorithm to deal with high-dimensional and complex data. The objective function of the new algorithm is modified with the help of two different entropy terms and a non-Euclidean way of computing the distance. The distance calculation formula enhances the efficiency of extracting the contribution of different features. The first entropy term helps to minimize the clusters' dispersion and maximize the negative entropy to control the clustering process, which also promotes the association between the samples. The second entropy term helps to control the weights of features since different features have different weights in the clustering process. Experiments on real-world datasets indicate that the proposed algorithm gives better clustering results than other algorithms. The experiments demonstrate the proposed algorithm's robustness by analyzing the parameters' sensitivity and comparing the computational distance formulas. In summary, the improved algorithm improves classification performance under noisy interference and high-dimensional datasets, increases computational efficiency, performs well in real-world high-dimensional datasets, and encourages the development of robust noise-resistant high-dimensional fuzzy clustering algorithms.

Tillage-Depth Verification Based on Machine Learning Algorithms

In an analysis of the penetration resistance and tillage depth of post-tillage soil, four surface-layer discrimination methods, specifically, three machine learning algorithms—Kmeans, DBSCAN, and GMM—and a curve-fitting method, were used to analyze data collected from the cultivated and uncultivated layers. Among them, the three machine learning algorithms found the boundary between the tilled and untilled layers by analyzing which data points belonged to which layer to determine the depth of the soil in the tilled layer. The curve-fitting method interpreted the intersection among data from the fitted curves of the ploughed layer and the un-ploughed layer as the tillage depth. The three machine learning algorithms were used to process a standard data set for model evaluation. DBSCAN’s discrimination accuracy of this data set reached 0.9890 and its F1 score reached 0.9934, which were superior to those of the other two algorithms. Under standard experimental conditions, the ability of DBSCAN clustering to determine the soil depth was the best among the four discrimination methods, and the discrimination accuracy reached 90.63% when the error was 15 mm. During field-test verification, the discriminative effect of DBSCAN clustering was still the best among the four methods. However, the soil blocks encountered in the field test affected the test data, resulting in large errors in the processing results. Therefore, the combined RANSCA robust regression and DBSCAN clustering algorithm, which can eliminate interference from soil blocks in the cultivated layer and can solve the problem of large depth errors caused by soil blocks in the field, was used to process the data. After testing, when the RANSCA and DBSCAN combined method was used to process all samples in the field and the error was less than 20mm, the accuracy rate reached 82.69%. This combined method improves the applicability of discrimination methods and provides a new method of determining soil depth.