scholarly journals Pure FPGA Implementation of an HOG Based Real-Time Pedestrian Detection System

Sensors ◽  
2018 ◽  
Vol 18 (4) ◽  
pp. 1174 ◽  
Author(s):  
Jian Luo ◽  
Chang Lin
Keyword(s):  
Real Time ◽  
Detection System ◽  
Pedestrian Detection ◽  
Image Sensor ◽  
Frame Rate ◽  
External Memory ◽  
Support Vector ◽  
Svm Classification ◽  
Prior Art ◽  
Memory Modules

In this study, we propose a real-time pedestrian detection system using a FPGA with a digital image sensor. Comparing with some prior works, the proposed implementation realizes both the histogram of oriented gradients (HOG) and the trained support vector machine (SVM) classification on a FPGA. Moreover, the implementation does not use any external memory or processors to assist the implementation. Although the implementation implements both the HOG algorithm and the SVM classification in hardware without using any external memory modules and processors, the proposed implementation’s resource utilization of the FPGA is lower than most of the prior art. The main reasons resulting in the lower resource usage are: (1) simplification in the Getting Bin sub-module; (2) distributed writing and two shift registers in the Cell Histogram Generation sub-module; (3) reuse of each sum of the cell histogram in the Block Histogram Normalization sub-module; and (4) regarding a window of the SVM classification as 105 blocks of the SVM classification. Moreover, compared to Dalal and Triggs’s pure software HOG implementation, the proposed implementation‘s average detection rate is just about 4.05% less, but can achieve a much higher frame rate.

scholarly journals A streaming multi-class support vector machine classification architecture for embedded systems

2019 ◽  
Vol 16 (3) ◽  
pp. 1286
Author(s):  
Jeevan Sirkunan ◽  
Jia Wei Tang ◽  
Nasir Shaikh-Husin ◽  
Muhammad Nadzir Marsono
Keyword(s):  
Embedded System ◽  
Test Data ◽  
Data Transfer ◽  
Pedestrian Detection ◽  
External Memory ◽  
Support Vector ◽  
Data Input ◽  
Maximum Throughput ◽  
Process Data ◽  
Svm Classification

<p>Pedestrian detection, face detection, speech recognition and object detection are some of the applications that have benefited from hardware-accelerated SVM. SVM classification computational complexity makes it challenging for designing hardware architecture with real-time performance and low power consumption. On an embedded streaming architecture, test data are stored on external memory and transferred in streams to the FPGA device. The hardware<br />implementation for SVM classification needs to be fast enough to keep up with the data transfer speed. Prior implementation throttles data input to avoid overwhelming the computational unit. This results in a bottleneck in overall streaming architecture as maximum throughput could not be obtained. In this work, we propose a streaming architecture multi-class SVM classification for embedded system that is fully pipelined and able to process data continuously with out any need to throttle data stream input. The proposed design is targeted for embedded platform where test data is transferred in streams from an external memory. The architecture was implemented on Altera Cyclone IV platform. Performance analysis on our proposed architecture is done with regards to the number of features and support vectors. For validation, the results obtained is compared with LibSVM. The proposed architecture is able to produce output rate identical to test data input rate.</p>

scholarly journals Low-Power Pedestrian Detection System on FPGA

Proceedings ◽  
2019 ◽  
Vol 31 (1) ◽  
pp. 35 ◽  
Author(s):  
Vinh Ngo ◽  
David Castells-Rufas ◽  
Arnau Casadevall ◽  
Marc Codina ◽  
Jordi Carrabina
Keyword(s):  
Low Power ◽  
Detection System ◽  
Low Cost ◽  
Pedestrian Detection ◽  
Support Vector ◽  
Embedded Processor ◽  
Car Industry ◽  
Field Programmable ◽  
On Chip ◽  
Histogram Of Gradients

Pedestrian detection is one of the key problems in the emerging self-driving car industry. In addition, the Histogram of Gradients (HOG) algorithm proved to provide good accuracy for pedestrian detection. Many research works focused on accelerating HOG algorithm on FPGA (Field-Programmable Gate Array) due to its low-power and high-throughput characteristics. In this paper, we present an energy-efficient HOG-based implementation for pedestrian detection system on a low-cost FPGA system-on-chip platform. The hardware accelerator implements the HOG computation and the Support Vector Machine classifier, the rest of the algorithm is mapped to software in the embedded processor. The hardware runs at 50 Mhz (lower frequency than previous works), thus achieving the best pixels processed per clock and the lower power design.

Real-Time Detection System of Soil Moisture Based on GPRS

2011 ◽  
Vol 464 ◽  
pp. 175-178
Author(s):  
Rong Biao Zhang ◽  
Jing Jing Guo ◽  
Qi Wang ◽  
Lei Zhang ◽  
Xian Lin Wang
Keyword(s):  
Soil Moisture ◽  
Real Time ◽  
Detection System ◽  
Field Tests ◽  
Support Vector ◽  
Calibration Model ◽  
Vector Machines ◽  
Moisture Detection ◽  
Software And Hardware ◽  
Seamless Connection

Real-time monitoring of soil moisture is essential for agricultural production. In this paper, an improved system is designed based on GPRS technology for real-time detecting soil moisture, a salinity calibration model is established based on Least Squares Support Vector Machines on MatLAB (LS-SVMlab) for improving detection precision. The transmission of soil moisture information is the key technology of the system, by software and hardware design we have solved the problems of data congestion, off-line, and moving the monitoring terminal at any time, which still restrict the application of GPRS in soil moisture detection. Field tests show that the system can realize seamless connection between the collection nodes and remote host, and acquire soil moisture accurately. Simultaneously, the time of re-networking has been shortened greatly.

Snow Detection using In-Vehicle Video Camera with Texture-Based Image Features Utilizing K-Nearest Neighbor, Support Vector Machine, and Random Forest

2019 ◽  
Vol 2673 (8) ◽  
pp. 221-232 ◽  
Author(s):  
Md Nasim Khan ◽  
Mohamed M. Ahmed
Keyword(s):  
Real Time ◽  
Prediction Accuracy ◽  
Nearest Neighbor ◽  
Detection System ◽  
Weather Conditions ◽  
Video Camera ◽  
Image Features ◽  
Support Vector ◽  
K Nearest Neighbor ◽  
Weather Information

Snowfall negatively affects pavement and visibility conditions, making it one of the major causes of motor vehicle crashes in winter weather. Therefore, providing drivers with real-time roadway weather information during adverse weather is crucial for safe driving. Although road weather stations can provide weather information, these stations are expensive and often do not represent real-time trajectory-level weather information. The main motivation of this study was to develop an affordable in-vehicle snow detection system which can provide trajectory-level weather information in real time. The system utilized SHRP2 Naturalistic Driving Study video data and was based on machine learning techniques. To train the snow detection models, two texture-based image features including gray level co-occurrence matrix (GLCM) and local binary pattern (LBP), and three classification algorithms: support vector machine (SVM), k-nearest neighbor (K-NN), and random forest (RF) were used. The analysis was done on an image dataset consisting of three weather conditions: clear, light snow, and heavy snow. While the highest overall prediction accuracy of the models based on the GLCM features was found to be around 86%, the models considering the LBP based features provided a much higher prediction accuracy of 96%. The snow detection system proposed in this study is cost effective, does not require a lot of technical support, and only needs a single video camera. With the advances in smartphone cameras, simple mobile apps with proper data connectivity can effectively be used to detect roadway weather conditions in real time with reasonable accuracy.

scholarly journals The Inspection of Raw- Silk Defects Using Image Vision

2018 ◽  
Vol 13 (3) ◽  
pp. 155892501801300 ◽  
Author(s):  
Qingtian Pan ◽  
Miao Chen ◽  
Baoqi Zuo ◽  
Yucai Hu
Keyword(s):  
Detection System ◽  
Image Sensor ◽  
Support Vector ◽  
Technological Parameters ◽  
Image Capture ◽  
Charge Coupled Device ◽  
The Moment ◽  
A Charge ◽  
Ccd Image

The inspection of defects is one of the most important aspects in the quality inspection of raw silk. We introduce a raw-silk defect detection system based on image vision and image analysis that is accurate and objective. In the experimental phase, we develop an image-acquisition section—which includes a charge-coupled device (CCD) image sensor, a telecentric lens, a light source, and a raw-silk winding device to capture the raw silk images steadily. After the image capture stage, an image-processing section tasked with threshold segmentation and morphology operations is carried out to obtain the defects of raw silk. To classify the raw-silk defects accurately and quickly, we propose an area method for the classification of raw-silk defects into five categories: larger defects, large defects, common defects, small defects, and smaller defects. Meanwhile, in order to recognize the common raw-silk defects—e.g., Bavella silk, nodes, and loose ends—that cannot be detected by the Uster evenness tester, the moment invariants of each segmented region of the images are extracted and used as the input of support vector machine(SVM).A SVM is designed as a classifier to recognize the samples. The experimental results show that the proposed method can recognize these common raw-silk defects effectively. According to the new classification and accurate recognition of raw-silk defects using the proposed method, we can improve the inspection standards for raw silk and advise raw-silk reeling enterprises seeking to optimize the technological parameters.

Sign in / Sign up

Export Citation Format

Share Document