Abstract
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.
Highlights
Real-time pedestrian detection is an important technology for modern society [1] in many applications, such as surveillance [2], intelligence vehicle systems [3], and robot navigation [4].Some studies have extracted diverse features from an image and have found appropriate classification methods to perform robust pedestrian detection [5,6,7,8,9]
The original histogram of oriented gradients (HOG) divided the orientations of gradients into several bins, and used the magnitudes of each gradient as the weight to generate the cell histogram in the Cell Histogram Generation sub-module
The distribution of the cell histogram is computed in the Block Histogram
Summary
Real-time pedestrian detection is an important technology for modern society [1] in many applications, such as surveillance [2], intelligence vehicle systems [3], and robot navigation [4].Some studies have extracted diverse features from an image and have found appropriate classification methods to perform robust pedestrian detection [5,6,7,8,9]. The histogram of oriented gradients (HOG) [10], was proposed by Dalal and Triggs in 2005. It is an efficient feature extraction algorithm, and it can accurately detect a pedestrian in difficult conditions, such as deformation, rotation, or illumination changes. This study implements the HOG feature extraction algorithm, and the SVM classification on a single FPGA. The detection window in the proposed scheme consists of 64 × 128 pixels in size, and it contains 7 × 15 blocks. The pedestrian detection process consists of two parts: First, it obtains the descriptors of each detection window by using the HOG feature extraction algorithm.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
