TRAFFIC VIOLATION DETECTION USING VIDEO PROCESSING

Reyna Talitha L. Santos

Ateneo De Zamboanga University

CS 322
Maureen B. Gallardo
January 31 2018

1
 INTRODUCTION

A. Background of the Study

Land Transportation Office (LTO) is one of the agency of the Philippine government

that is responsible for all land transportations, registration of new vehicles, issuance of

driver’s license and enforcing traffic laws in the country. They rationalize the land

transportation services and facilities as stated in their mission. Promoting safety and

comfort in land travel is a continuing commitment of the said agency.

According to the LTO- Zamboanga City Branch Administration and Registration

Department, there are 81, 232 number of vehicles registered in the said City which

includes 1,788 jeepneys, 73,264 PUV, and 4,572 tricycles for the year of 2017 which are

commonly used in the said city.

The sudden rise of motor accidents from 2010

2
B. Statement of the Problem

The focal problem of the study centers around motorcycle rider casualties due to

failure to wear the standard protective motorcycle helmet or failure to require the back

rider to wear the standard protective motorcycle helmet under Republic Act 10054. Being

involved in a motorcycle crash can cause numerous injuries; this is true whether a rider

has a helmet on or not. Kovacic (2013) states, helmets are estimated to be 37 percent

effective in preventing fatal injuries to motorcyclists. Unfortunately, not wearing a helmet

increases the chances of much more serious injuries and the severity of the following:

Head injuries cause some of the most severe damage a person’s body can experience,

and sadly, they’re almost always preventable. Motorcycle helmets are meant to take the

brunt of the force when a biker’s head hits the pavement. Studies have shown that simply

walking into something solid can cause a skull fracture, then hitting the pavement at 55

MPH can do to a human skull without a helmet. Severe head injuries can cause brain

damage, coma or death. Spinal injuries are one of the most feared injuries in the world,

and not wearing a helmet when on a motorcycle can increase the risk of experiencing

one. Helmets not only protect a biker’s head; they help keep their head stationary. If a

person’s head remains stationary during an accident, their chances of not sustaining a

serious neck or spinal injury increases. Unfortunately, those who do experience traumatic

spinal injuries can face partial paralysis, full on paraplegia or even death. Serious Facial

Abnormalities. Unfortunately, even riders who don’t suffer serious brain or spinal injuries

can have their faces completely destroyed by not wearing a helmet.

Currently there are 4,872,065 registered motorcycles in total in the Philippines in

2016 but how often do motorcycle accidents and casualties occur? Are they frequent

enough to be considered a problem? Based on records on the nationwide tally for daily
3
transport incidents in 2015 from the Philippine Statistics Authority, 69% of the of the total

identified transport incidents nationwide are motorcycle accidents and casualties.

Motorcycle crashes remained the leading cause of death and injury on Metro Manila

roads, a government traffic accident recording and analysis report showed. The 2015

Metro Manila Accident Recording and Analysis System report showed that motorcycles

have the “highest fatality accident rate” with 262 the total number of deaths, followed by

trucks (129) and private cars (125). The report noted that the number of road accidents in

Metro Manila has increased to 95,615 incidents in 2015, up by 5,357 from the 90,258

cases recorded in 2014. From 6,244 injuries in 2010, the number of recorded injuries

increased to 19,852 in a span of 5 years. In addition, each quarterly DOH report notes the

road traffic-related injuries involving motorcycles, 98.7% of fatalities involved motorists

without helmets, and 1.3% of fatalities involved motorists wearing a helmet.

4
C. Objectives

General Objectives

To develop a video processing system that detects if motorcycle riders have failed

to wear the standard protective motorcycle helmet.

Specific Objectives

a. To create a module that detects the human head using Head Region

Detection techniques

b. To create a module that detects plate numbers of the vehicle that violates

using plate number optical character recognition;

c. To create a Traffic Violation Information System that can saves and

retrieves video footages of the violation, date and time of violation, plate

number, vehicle type, color and brand of the vehicle and its case status.

5
D. Conceptual Framework

Figure1. Conceptual Framework

To achieve the general objectives of the study, the system is decomposed to

different modules. First, the system will first view video footages of all vehicles and will

detect if there are that using image subtraction. Then it will check if there are violations

violated by that vehicle in the violation detection module. In this module, detects for heads

that do not have helmets on as a basis of violation. If the system is able to detect a

violation, it will then save the footage of that part of that video. Then at a certain time, it

will post-process the video footages by recognizing its plate number through optical

character recognition and other information which will then be saved in the Traffic

Violation Information System.

6
E. Scope and Limitation

The study will focus on developing video processing system the violation of RA

10054 or the failure to wear the standard protective helmet with the use of video

processing for the City of Zamboanga and for LTO Regional Office Region IX. The scope

involves the monitoring of motorcyclists on a 24 hours basis only in major intersections

that are known to have high volumes of vehicles that flow throughout the day. The video

processing system also includes Traffic Violation Information System that saves the video

footage date and time of the occurrence and plate number of the violator. Since we only

aim to detect the violation of RA 10054 then the kind of violation by default will be set to

said violation. The system does not save all video footages 24 hour on the Traffic Violation

Information System. The system will choose to save footages that only contain violations

and deletes the rest otherwise.

The system can only detect the violation of RA 10054 and has its primary focus on

monitoring two wheeled motorcycles excluding other types of motorcycle variants such

as tricycles and the two-wheeled tractor (also known as “Kuliglig”).

7
F. Significance of the Study

The development of this system will enhance road safety and deter fatal motorcycle

involved casualties. The effectivity of the system will help improve the implementation of

the RA 10054 as the selected road will be monitored on a 24 hour basis every day.

8
 REVIEW OF RELATED LITERATURE

1. Republic Act No. 10054 - "Motorcycle Helmet Act of 2009"

The declared law imposes the mandatory implementation of the Republic Act to all

motorcycle riders, including drivers and back riders, shall at all times wear standard

protective motorcycle helmets while driving, whether long or short drives, in any type of

road and highway. Standard protective motorcycle helmets are appropriate types of

helmets for motorcycle riders that comply with the specifications issued by the Department

of Trade and Industry (DTI). The DTI has issued guidelines, which should include the

specifications regarding standard protective motorcycle helmets. The DTI, through the

BPS, shall conduct a mandatory testing of all manufactured and imported motorcycle

helmets in the Philippines. All manufacturers and importers of standard protective

motorcycle helmets are required to secure a PS license or ICC prior to the sale and

distribution of their products, only those standard protective motorcycle helmets bearing

the PS or ICC mark shall be sold in the market.

2. Land Transportation Office’s process on penalizing violators of RA 10054

Any person caught not wearing the standard protective motorcycle helmet in

violation of this Act shall be punished with a fine of One thousand five hundred pesos

(Php1,500.00) for the first offense; Three thousand pesos (Php3,000.00) for the second

offense; Five thousand pesos (Php5,000.00) for the third offense; and Ten thousand

pesos (Php10,000.00) plus confiscation of the driver's license for the fourth and

succeeding offenses.

9
3. System Peripherals

Our system heavily relies on video processing as a means to detect the helmet

violation. The following peripherals are needed for the system.

3.1. Smart Camera

According to Belbachir (2009), a smart camera or intelligent camera is a vision

system which, in addition to image capture circuitry, is capable of extracting application

specific information from the captured images, along with generating event descriptions

or making decisions that are used in an intelligent and automated system. A smart camera

is a self-contained, standalone vision system with built-in image sensor in the housing of

an industrial video camera. The purpose of the smart camera is to capture footages for

the system that are to be processed later on.

3.2. IP Camera

An Internet protocol camera, or IP camera, is a type of digital video camera

commonly employed for surveillance, and which, unlike analog closed-circuit television

(CCTV) cameras, can send and receive data via a computer network and the Internet.

Although most cameras that do this are webcams, the term IP camera or netcam is usually

applied only to those used for surveillance that can be directly accessed over a network

connection. The usage of the IP camera is similar with the smart camera as

aforementioned.

3.3. Wireless Connectivity

A wireless network is a computer network that uses wireless data connections

between network nodes. Examples of wireless networks include cell phone networks,

wireless local area networks (WLANs), wireless sensor networks, satellite communication
10
networks, and terrestrial microwave networks. This will enable communication between

the peripherals and the main violation detection module and information system

3.4 Server

In computing, a server is a computer program or a device that provides

functionality for other programs or devices, called "clients". This architecture is called the

client–server model, and a single overall computation is distributed across multiple

processes or devices. Servers can provide various functionalities, often called "services",

such as sharing data or resources among multiple clients, or performing computation for

a client. A single server can serve multiple clients, and a single client can use multiple

servers

4. Algorithms

There have been numerous studies on differencing the background from the image and

identifying features such as elliptical arcs that are helpful in the detection of helmets. The first and

most important step is the detection of the human head and the helmet with the following

algorithms.

4.1 Background Subtraction Using Time Differencing

Moving objects in a static scene can be detected by applying the Background

Subtraction Algorithms. It first subtracts consecutive frames from a footage then it leaves

only moving objects that have been detected.

11
4.3 Blob Analysis

According to Gill (2017), blob Analysis is a fundamental technique of machine

vision based on analysis of consistent image regions. As such it is a tool of choice for

applications in which the objects being inspected are clearly discernible from the

background. Diverse set of Blob Analysis methods allows to create tailored solutions for

a wide range of visual inspection problems.

The basic scenario of the Blob Analysis solution consists of the following steps:
1. Extraction - in the initial step one of the Image Thresholding techniques is

applied to obtain a region corresponding to the objects (or single object) being

inspected.

2. Refinement - the extracted region is often flawed by noise of various kind (e.g.

due to inconsistent lightning or poor image quality). In the Refinement step the

region is enhanced using region transformation techniques.

3. Analysis - in the final step the refined region is subject to measurements and the

final results are computed. If the region represents multiple objects, it is split into

individual blobs each of which is inspected separately.

4.4 Optical Character Recognition

According to Schantz (1862), optical character recognition is the conversion of

images of any typed text, handwritten or printed into machine-encoded text. These

images are usually retrieved from documents, scene-photo or from subtitle

superimposed on an image. It is widely used as a form of information entry. It is a

12
common method of digitizing printed texts so that they can be digitally or electronically

edited and searched.

4.5 Information System

An information system (IS) is an organized system for the collection, organization,

storage and communication of information. More specifically, it is the study of

complementary networks that people and organizations use to collect, filter, process,

13
create and distribute data. This will be needed for storing the violator information such

as footage, date and time of occurrence, the plate number and if the case is pending or

has been resolved.

5. Related Studies

5.1. Catch-All

CATCH-ALL is a complex system that captures violators on a 24-hour basis using

artificial intelligence that does not need any intervention or manual manipulation from

humans. The specialized CCTV cameras have image and video capturing capabilities

through fiber optic technology. The images are then processed in a laboratory situated

within DLSU containing multiple computers. Features include, identifying the car,

profiling the car whether it is small, medium, or large, what type of car it is, identifying

the motions of the car from one location to another. Car tracking, movement monitoring

and identifying car speed. Once the plate numbers are stored, the system can clip the

video for storage and documentation purposes. Should the owner have future

complaints, the video clip will be presented as proof that the corresponding vehicle has

violated traffic rules.

14
 METHODOLOGY

The study has proposed a vision-based detection of non-observance of the RA

law thus this will be divided into three major stages. The detection and classification,

detection of helmet use and the case opening.

1. Detection and Classification

The first stage, detection and classification must first determine if there are any

moving objects in a static scene in a footage. Having detected objects in motion, the

system then classifies them if it’s a four-wheeled vehicle or a motorcycle. In this study,

the system would disregard non-motorcycle vehicles.

1.1 Detecting static scene or the background

The main objective of this step is to obtain an image that will be used to detect

moving objects. Any form of moving object is considered a possible vehicle that will be

classified later on.

15