Sorting Accuracy Research Articles

Fast continuous streaming sort in big streaming data environment under fixed-size single storage.

Big streaming data environment concerns a complicated scenario where data to be processed continuously flow into a processing unit and certainly cause a memory overflow problem. This obstructs the adaptation of deploying all existing classic sorting algorithms because the data to be sorted must be entirely stored inside the fixed-size storage including the space in internal and external storage devices. Generally, it is always assumed that the size of each data chunk is not larger than the size of storage (M) but in fact the size of the entire stream (n) is usually much larger than M. In this paper, a new fast continuous streaming sorting is proposed to cope with the constraint of storage overflow. The algorithm was tested with various real data sets consisting of 10,000 to 17,000,000 numbers and different storage sizes ranging from 0.01n to 0.50n. It was found that the feasible lower bound of storage size is 0.35n with 100% sorting accuracy. The sorting time outperforms bubble sort, quick sort, insertion sort, and merge sort when data size is greater than 1,000,000 numbers. Remarkably, the sorting time of the proposed algorithm is 1,452 times less than the sorting time of external merge sort and 28.1767 times less than the sorting time of streaming data sort. The time complexity of proposed algorithm is O(n) while the space complexity is O(M).

Measuring Shape Parameters of Pearls in Batches Using Machine Vision: A Case Study.

To solve the problem of low precision of pearl shape parameters’ measurement caused by the mutual contact of batches of pearls and the error of shape sorting, a method of contacting pearls’ segmentation based on the pit detection was proposed. Multiple pearl images were obtained by backlit imaging, the quality of the pearl images was improved through appropriate preprocessing, and the contacted pearl area was extracted by calculating the area ratio of the connected domains. Then, the contour feature of the contact area was obtained by edge tracking to establish the mathematical model of the angles between the edge contour points. By judging the angle with a threshold of 60° as the candidate concave point, a concave point matching algorithm was introduced to get the true concave point, and the Euclidean distance was adopted as a metric function to achieve the segmentation of the tangent pearls. The pearl shape parameters’ model was established through the pearl contour image information, and the shape classification standard was constructed according to the national standard. Experimental results showed that the proposed method produced a better segmentation performance than the popular watershed algorithm and morphological algorithm. The segmentation accuracy was above 95%, the average loss rate was within 4%, and the sorting accuracy based on the shape information was 94%.

Research on intelligent optimization separation technology of coal and gangue base on LS-FSVM by using a binary artificial sheep algorithm

The separation of coal gangue is affected by various complicated factors. Noise, complex texture characteristics and sample isolation seriously affect the extraction of gangue. In order to improve the accuracy and stability of coal gangue separation, a coal gangue separation method based on artificial binary sheep algorithm optimized hyperplane membership fuzzy least squares support vector machine comprehensive feature classifier (BASA-LS-FSVM) was proposed. The parameters of gray mean, gray frequency and entropy, energy, contrast and correlation were extracted from the gray matrix, and the parameters of fuzzy support vector machine classifier were optimized by artificial sheep binary algorithm. The improved BASA-LS-FSVM classification model was trained under the same sample. In order to verify the superiority of the model and algorithm, three comparison models are constructed: single feature classifier model, fusion feature classifier model and comprehensive feature classifier model. The synthetic feature classifier uses gray scale, texture and fusion features as the feature classifier model of normal plane least squares vector machine optimized by BASA, and trains the model under the same training samples. The training results show that the training correlation under the monomer characteristics has a better effect on the classification of characteristic quantity, and the sorting accuracy of coal gangue is 90% and 91%, respectively. The combined characteristic parameters of training frequency and ash have better separation effect, and the separation accuracy of gangue is higher. Under the comprehensive characteristics, the separation accuracy of coal gangue can reach 98% and 99%. The results show that the sorting effect is the best under the comprehensive characteristics. Compared with other models, the fitness function value of BASAA-NP-FSVM model reached the optimal value in the 48th generation, and the separation accuracy of coal gangue reached 98% and 99%, respectively.

Radar Emitter Classification Based on a Multiperspective Collaborative Clustering Method and Radar Characteristic Spectrum

Modern cognitive electronic reconnaissance methods for radar systems must contend with the complex electromagnetic environments arising from the deployment of multiple signal sources and radar countermeasures, which greatly limit access to the degree of prior information required to enable effective target recognition. The present work addresses this issue by proposing a multiperspective collaborative clustering method for sorting radiation sources based on the multiperspective information of radar signals. In contrast to conventional collaborative training approaches, which are suitable only for semisupervised learning, the proposed multiperspective collaborative clustering method performs unsupervised clustering, cluster label transfer, and dimensionality reduction by linear discriminant analysis iteratively based on the differences between the clustering results obtained from two signal perspectives radiation signal sorting can be conducted in a noncooperative context. The results of comparative experiments demonstrate that the proposed multiperspective sorting method can make full use of the difference information between basic signal characteristics and intrapulse features and thereby improve the accuracy of clustering-based radiation source sorting. Accordingly, the sorting ability of the proposed method is superior to those of other state-of-the-art clustering methods and that of the single-perspective clustering-based sorting method.

Robotics for control strategy optimization and simulation in automatic cigarette sorter

In the world of advancement and rapid growth, new retail marketization and distribution require a substantial degree of versatility and resilience in the manufacturing plant to accommodate demand transitions. A Digital Matching Structure (DMS) is built such that real-time data from the framework for processing the physical goods can be aggregated and then translated into the digital context. A collaborative design approach is implemented in the DMS to automate stacked cigarette labelling and sorting. The periodic optimal decisions can be achieved from the joint optimization model by obtaining online details from the specific unit for the management network and then feeding them to the quasi-physical simulation platform in the DMS for validation of results. A sorting robot built-in DMS can be configured to catch and sort cigarettes on a conveyor belt to ensure high performance and eliminate the possibility of manual sorting. Digital Matching Structure – Sorting Robot (DMS-SR) is the solution recommended for the proposed integration. This article explores how finely packaged cigarettes are handled on the conveyor belt and how the packaging line works to understand and filter the product. The proposed integration method is assessed using appropriate mathematical models for validating its consistency and verifying the accuracy of sorting. Eventually, the proposed autonomous system parameters have been enhanced by tests on a real machine to boost the sorting performance. It was found that better sorting could be obtained when the conveyor speed of around 0.36 m/s. When there is sufficient load, growing the momentum and shortening the declaration/accelerationtime will produce the optimum efficacy. The DMS-SR currently has a sorting output of about 1900 cigarettes per hour, highly precise.

Nondestructive Dimension Sorting by Soft Robotic Grippers Integrated with Triboelectric Sensor.

In smart logistics, traditional manual sorting and sorting systems based on rigid manipulators limit the warehousing development and damage the goods. Here, a nondestructive sorting method based on bionic soft fingers is proposed. This method is implemented by the soft robotic gripper (SRG) for grasping of the breakable objects, the triboelectric sensor (TES) for size sorting of the objects, and the signal processing module. In the fabrication of SRG, the silicon rubber is prepared by controlling the material synthesis process, and its Young's modulus is 600.91 kPa, which is comparable to the Young's modulus of skin tissue. Also, the maximum input pressure of SRG is 71.4 kPa. The TES has a linear relationship between pulse number and sliding displacement, and its resolution is 3 mm. It induces pulse signal sequences to quantify the SRG bending state and thus realize the size sorting of objects. Additionally, a nondestructive sorting system based on TES and SRG has been developed for fruit sorting (e.g., apples, oranges), enabling nondestructive grasping and accurate sorting. Its sorting range is 70-120 mm, and the sorting accuracy rate is up to 95%. This work also provides a way for the application of SRG and triboelectric sensors in the sorting field.

Soft clustering of retired lithium-ion batteries for the secondary utilization using Gaussian mixture model based on electrochemical impedance spectroscopy

Rapid sorting and reasonable regrouping of retired lithium-ion batteries (LIBs) are directly related to the economy and safety of the second-life utilization. However, the efficiency and accuracy of sorting for the retired LIBs needs to be further improved, and the regrouping method is still in the exploratory stage. In this study, a soft clustering method based on the Gaussian mixture model (GMM) using electrochemical impedance spectroscopy (EIS) is proposed to address these issues. In this method, the multi-dimensional clustering criteria are extracted from EIS, and the capacity is quickly estimated based on the EIS using a neural network. Furthermore, the ageing factors of six criteria are constructed to realize the soft clustering of retired cells corresponding to three ageing modes. The simulation results show that it only takes about 10 min to obtain the capacity of each cell, and the error is within 4%. Moreover, the clustering probability of each cell under different ageing modes is obtained using GMM, which is useful for flexible grouping of cells. Finally, the proposed methods are evaluated by experiments, and results show that the consistency of the regrouped cells using the proposed soft-clustering method is nearly doubled than that of the random regrouped cells.

Objective evaluation of stimulation artefact removal techniques in the context of neural spike sorting

Objective. We present a framework to objectively test and compare stimulation artefact removal techniques in the context of neural spike sorting. Approach. To this end, we used realistic hybrid ground-truth spiking data, with superimposed artefacts from in vivo recordings. We used the framework to evaluate and compare several techniques: blanking, template subtraction by averaging, linear regression, and a multi-channel Wiener filter (MWF). Main results. Our study demonstrates that blanking and template subtraction result in a poorer spike sorting performance than linear regression and MWF, while the latter two perform similarly. Finally, to validate the conclusions found from the hybrid evaluation framework, we also performed a qualitative analysis on in vivo recordings without artificial manipulations. Significance. Our framework allows direct quantification of the impact of the residual artefact on the spike sorting accuracy, thereby allowing for a more objective and more relevant comparison compared to indirect signal quality metrics that are estimated from the signal statistics. Furthermore, the availability of a ground truth in the form of single-unit spiking activity also facilitates a better estimation of such signal quality metrics.

Deep Vision Servo Hand-Eye Coordination Planning Study for Sorting Robots

Within the context of large-scale symmetry, a study on deep vision servo multi-vision tracking coordination planning for sorting robots was conducted according to the problems of low recognition sorting accuracy and efficiency in existing sorting robots. In this paper, a kinematic model of a mobile picking manipulator was built. Then, the kinematics design of the orwX, Y, Z three-dimensional space manipulator was carried out, and a method of deriving and calculating the base position coordinates through the target point coordinates, the current moving chassis center coordinates and the determined manipulator grasping attitude conditions was proposed, which realizes the adjustment of the position and attitude of the moving chassis as small as possible. The multi-vision tracking coordinated sorting accounts 79.8% of the whole cycle. The design of a picking robot proposed in this paper can greatly improve the coordination symmetry of logistic package target recognition, detection and picking.

Efficient Recognition and Automatic Sorting Technology of Waste Textiles Based on Online Near infrared Spectroscopy and Convolutional Neural Network

In order to better recycle waste textiles and save resources, intelligent identification and sorting equipment and technology are urgently needed. In this work, an online near infrared (NIR) spectral library was established by utilizing self-developed online NIR device, including polyester, cotton, wool, silk, viscose, nylon, acrylic, polyester/cotton, polyester/wool, polyester/nylon, polyester/viscose, nylon/spandex and silk/cotton. Importantly, artificial intelligence technology was introduced into the identification and sorting of waste textiles, and two online NIR qualitative identification models covering above 13 kinds of waste textiles were constructed by the convolutional neural network (CNN) and Baidu deep learning platform PaddlePaddle. First, the input one-dimensional spectral data (901-2500 nm) was normalized and converted into a two-dimensional grayscale image of 40*40 pixels. Then feature extraction, compression and dimension reduction of multiple spectra were carried out through convolution and pooling. Finally, the category probability value of each kind of waste textiles was calculated by the CNN model and the maximum value was taken as the final classification of the fabric. Online identification tests were performed using 526 samples as an external validation set, presenting an accuracy of two CNN qualitative identification models were both more than 95.4%. In addition, the accuracy of online identification and sorting was above 95%, and the recognition and sorting time of each sample is less than 2 s, which can perform the efficient identification and automatic sorting of waste textiles.

Fluorescence lifetime activated droplet sorting (FLADS) for label-free sorting of Synechocystis sp. PCC6803.

This study presents the label-free sorting of cyanobacterial cells in droplets with single-cell sensitivity based on their fluorescence lifetime. We separated living and dead cyanobacteria (Synechocystis sp. PCC6803) using fluorescence lifetime signals of the photopigment autofluorescence to indicate their photosynthetic activity. We developed a setup and a chip design to achieve live/dead sorting accuracies of more than 97% at a droplet frequency of 100 Hz with a PDMS-based chip system and standard optics using fluorescence lifetime as the sorting criterion. The obtained sorting accuracies could be experimentally confirmed by cell plating and observing the droplet sorting process via a high-speed camera. The herein presented results demonstrate the capabilities of the developed system for studying the effects of stressors on cyanobacterial physiology and the subsequent deterministic sorting of different stress-response phenotypes. This technology eliminates the need for tedious staining of cyanobacterial cells, which makes it particularly attractive for its application in the field of phototrophic microbial bio(techno)logic and in the context of cell secretion studies.

Research on lithium battery sorting method based on image adaptive recognition of feature points

When sorting batteries in factories, fixed feature points in a discharge voltage platform are usually used as a sorting basis with expert experience, causing poor accuracy in battery sorting. This study proposes a method based on feature point image adaptive recognition (FPIAR) and combined a semi-supervised fuzzy C-means (SFCM) clustering algorithm to achieve battery sorting. Accordingly, FPIAR obtained the multi-dimensional feature coordinate dataset by adaptively identifying the discharge voltage curve of each battery. The centre point of the accumulation of coordinates in the dataset was found by mean-shift algorithm and was used as the feature point of battery sorting voltage. In addition, the SFCM clustering algorithm was used to process the feature point of the sorting voltage to sort the batteries into groups. The simulation results and tests indicated that the SFCM clustering algorithm based on FPIAR processing can improve the accuracy of battery sorting.

Stiffness Assessment and Lump Detection in Minimally Invasive Surgery Using In-House Developed Smart Laparoscopic Forceps.

Minimally invasive surgery (MIS) incorporates surgical instruments through small incisions to perform procedures. Despite the potential advantages of MIS, the lack of tactile sensation and haptic feedback due to the indirect contact between the surgeon’s hands and the tissues restricts sensing the strength of applied forces or obtaining information about the biomechanical properties of tissues under operation. Accordingly, there is a crucial need for intelligent systems to provide an artificial tactile sensation to MIS surgeons and trainees. This study evaluates the potential of our proposed real-time grasping forces and deformation angles feedback to assist surgeons in detecting tissues’ stiffness. A prototype was developed using a standard laparoscopic grasper integrated with a force-sensitive resistor on one grasping jaw and a tunneling magneto-resistor on the handle’s joint to measure the grasping force and the jaws’ opening angle, respectively. The sensors’ data are analyzed using a microcontroller, and the output is displayed on a small screen and saved to a log file. This integrated system was evaluated by running multiple grasp-release tests using both elastomeric and biological tissue samples, in which the average force-to-angle-change ratio precisely resembled the stiffness of grasped samples. Another feature is the detection of hidden lumps by palpation, looking for sudden variations in the measured stiffness. In experiments, the real-time grasping feedback helped enhance the surgeons’ sorting accuracy of testing models based on their stiffness. The developed tool demonstrated a great potential for low-cost tactile sensing in MIS procedures, with room for future improvements. Significance: The proposed method can contribute to MIS by assessing stiffness, detecting hidden lumps, preventing excessive forces during operation, and reducing the learning curve for trainees.

Deep ToA Mask Based Recursive Radar Pulse Deinterleaving

In a complex electromagnetic environment, multiple radar signals of various modes are densely interleaved. In this environment, radar parameters overlap seriously and change continuously over time. Traditional radar pulse deinterleaving algorithms face severe challenges such as parameters missing, pulse jitter, and the increasing number of electronic countermeasure devices. In this paper, we propose a recursive deinterleaving algorithm based on blind signal separation and deep learning to cope with such a situation. The Recursive Deinterleaving Network (RDN) of Deep ToA Mask (DTM) maps the ToA train to a suitable feature space first. ToA coefficient masks of each radar emitter are estimated with the local and global context information of the radar pulse feature. Then the RDN sorts out several radar pulse trains recursively with the help of dual-path attention. It also predicts the number of emitters with nearly 100% accuracy and handles the unknown Pulse Repetition Interval (PRI) situation. More accurate pulse deinterleaving results can be obtained if the DTM utilizes more radar parameters through proper pre-processing fine-tuning and post-processing re-clustering. The processing steps of the DTM are introduced in detail. The simulation shows it can achieve 97% sorting accuracy for multi-pulse interleaved radar train with jitter PRI and pulse missing. The DTM algorithm can also deal with the interleaved radar signals of different PRI modulations by re-clustering with noisy PDW information. On the premise of knowing the modulation type or PRI information, the pulse train deinterleaving accuracy of multi-modulation emitters is higher.

Image-based rock mixing ratio estimation by using illuminance analysis in underground mining

ABSTRACT Illuminance is a key factor in optical sorting in underground coal mining, since it affects the image grayscale and texture features, especially for objects that are sensitive to light, such as bright coal. Appropriate illuminance is beneficial for improving the accuracy of image-based rock mixing ratio estimation, which is essential for optical sorting in underground coal mining. In this study, an image acquisition system with controllable illuminance was established. Groups of images were obtained under six illuminance conditions: 3180, 10780, 17730, 24200, 30700, and 35600 lx. MATLAB code was developed to segment an image and extract image features from each segmented subregion. The relations between illuminance and the image features for bright coal were quantified. The differences in the features between bright coal and sandstone were discussed. Two theoretical models, including area model and volume model, were used for estimating rock mixing ratio of the digital image. Our main conclusions are as follows. In general, the image features of bright coal blocks are different from those of sandstone blocks. There were significant differences between bright coal and sandstone for the variance under 24200 lx, the kurtosis under 10780 lx, and the fractal dimension under 3180 lx. Therefore, the proper illuminance is necessary for accurately determining the image features of bright coal and could even moderately increase the differences in the image features between bright coal and sandstone. Rock mixing ratio obtained by the area model is in better agreement with the manual results than volume model in this study. This research is of great significance as it presents a novel idea for improving the accuracy of optical sorting for underground mining.

Study on flexible deformation of waste plastic film by fictitious particle method

Waste plastic sorting and recycling is an efficient and energy-saving way of domestic waste treatment. The deformation of waste plastic film affects the accuracy of plastic sorting. In this paper, a new method for calculating the deformation of waste plastic film, fictitious particle method, is proposed. The calculation model of film deformation is established and the formula of line element is deduced. The deformation law of plastic film is studied and the calculation results of fictitious particle method are obtained. The experimental platform was set up to obtain the film deformation data. The two-way fluid–solid coupling method was used for simulation experiment. Compared with the experimental results and simulation results, it is verified that the algorithm is in good agreement. The average error rate of the fictitious particle method is 4.26%, and the calculation time is 3.5 s. The algorithm has small amount of calculation and fast calculation speed. It laid the foundation for effectively improving the accuracy of wind separation and the optimization of plastic film wind separation machinery. • A new numerical calculation method called fictitious particle method is proposed. • The deformation law of the film in the separation process of waste plastic film was explored. • The fictitious particle method reduces the calculation amount and the calculation speed is fast. • The superiority of fictitious particle method in studying the deformation of waste plastic film was verified. • Improve the accuracy of wind sorting waste plastic film.

Stratified Sorting Method of Battery Module Considering SOH in Echelon Utilization

To improve the efficiency of large-scale retired power battery (RPB) sorting in echelon utilization, and a stratified sorting method is proposed in the paper which takes battery module (BM) as object. Firstly, the correlation characteristic between the single battery and BM is built, and the main parameters of affecting the attenuation characteristic of the BM are defined. According to the consistency of the static initial parameters of the battery module, the first sorting is accomplished. Secondly, the characteristics correlation function between the single battery and the BM is established. The output characteristics of BM are calculated by using the equilibrium control method. Finally, the coupling relationship of the state of charge (SOC), depth of discharge, power characteristics and the state of health (SOH) of BM are established considering the SOH of BM comprehensively. The BM is sorted for the second time by using k-means clustering. The proposed sorting method is verified by simulation which effectively improves the accuracy of sorting and improves the battery's SOH.

The impact of reducing signal acquisition specifications on neuronal spike sorting.

Measuring electrical potentials in the extracellular space of the brain is a popular technique because it can detect action potentials from putative individual neurons. Electrophysiology is undergoing a transformation where the number of recording channels, and thus number of neurons detected, is growing at a dramatic rate. This rapid scaling is paving the way for both new discoveries and commercial applications; however, as the number of channels increases there will be an increasing need to make these systems more power efficient. One area ripe for optimization are the signal acquisition specifications needed to detect and sort action potentials (i.e., "spikes") to putative single neuron sources. In this work, we take existing recordings collected using Intan hardware and modify them in a way that corresponds to reduced recording performance. The accuracy of these degraded recordings to spike sort using MountainSort4 is evaluated by comparing against expert labels. We show that despite reducing signal specifications by a factor of 2 or more, spike sorting accuracy does not change substantially. Specifically, reducing both sample rate and bit depth from 30 kHz and 16 bits to 12 kHz and 12 bits resulted in a 3% drop in spike sorting accuracy. Our results suggest that current neural acquisition systems are over-specified. These results may inform the design of next generation neural acquisition systems enabling higher channel count systems.

Selective Single-Cell Sorting Using a Multisectorial Electroactive Nanowell Platform.

Current approaches in targeted patient treatments often require the rapid isolation of specific rare target cells. Stream-based dielectrophoresis (DEP) based cell sorters have the limitation that the maximum number of sortable cell types is equivalent to the number of output channels, which makes upscaling to a higher number of different cell types technically challenging. Here, we present a microfluidic platform for selective single-cell sorting that bypasses this limitation. The platform consists of 10 000 nanoliter wells which are placed on top of interdigitated electrodes (IDEs) that facilitate dielectrophoresis-driven capture of cells. By use of a multisectorial design formed by 10 individually addressable IDE structures, our platform can capture a large number of different cell types. The sectorial approach allows for fast and straightforward modification to sort complex samples as different cell types are captured in different sectors and therefore removes the need for individual output channels per cell type. Experimental results obtained with a mixed sample of benign (MCF-10A) and malignant (MDA-MB-231) breast cells showed a target to nontarget sorting accuracy of over 95%. We envision that the high accuracy of our platform, in addition to its versatility and simplicity, will aid clinical environments where reliable sorting of varying complex samples is essential.

UAV swarm based radar signal sorting via multi-source data fusion: A deep transfer learning framework

Traditional clustering algorithms can be applied for the pre-sorting step of radar signal sorting. It can effectively dilute the pulse stream and prevent the dense pulse stream from interfering pulse repetition interval (PRI) extraction. However, the pre-sorting deviation will cause interference and missing pulses during the main sorting process. To solve this problem, we deploy the unmanned aerial vehicle (UAV) swarm to monitor reconnaissance areas and put forward a novel deep transfer learning based signal sorting method. The UAV swarm can collect the pulses from different time and spatial domains, and interference and missing pulses in main sorting processing can be relieved dramatically. In our model, we pre-train our model with the data collected from multiple source areas, which corresponds to different areas detected by different parts of UAV swarms. Then we fine-tune our model with the data of the target area. The experimental results prove that the signal sorting accuracy of methods based on deep transfer learning, i.e., YOLO-MobileNet, F-RCNN and cascade RCNN, are higher than that of the baseline methods. In addition, the signal sorting accuracy of traditional methods based on deep learning can be greatly improved with the help of transfer learning.