International Journal of Engineering Science Invention Research & Development; Vol. IV, Issue VI, DECEMBER 2017 www.ijesird.com, E-ISSN: 2349-6185 A REVIEW OF RAINDROPLET DETECTION AND REMOVAL TECHNIQUES 1 1,2 K Murali Gopal, 2Ranjit Patnaik Assistant Professor, Department of Computer Science and Engineering, GIET, Gunupur (Odisha) – 765022 1 kmgopal@giet.edu, 2 ranjit.patnaik83@gmail.com Abstract—Bad weather condition decrease the surveillance video and the driving assistance system efficiency and accuracy. The impact of rain drop in the single images can make it difficult to distinguish the objects. Furthermore, a high quality single image is needed in numerous areas such as in object recognition and detection noise removal and weather condition removal. Rainy weather outdoor images and videos reduce the visibility, performance of computer vision algorithms, which use for extracting features and information from images. This paper will present a review of restoration raindrop detection and removal from single image which has different techniques of used in video. Keywords— raindrop, bad weather, Rain drop removial,. I. INTRODUCTION Image processing and computer vision research has a great history where many areas are addressed [1] like image compression, object detection and its performance, enhancement in image in many domains like medical, industrial, surveillance and weather forecasting [2]. Aim of this paper is to summarize the classification bad weather in computer vision, types of noise created by rain drops, and its removal technique& its performance. Before discussing about the rain drop and bad weather, let us discuss the image enhancement by removing the noise. Depending on the requirement the image enhancement can treat an image to increase the quality of the image through removing the blurred, noise or balancing the contrast or brightness. In 1969, Huage [3] describe the image enhancement parameters likecrispening, contrast enhancement, noise removal and inverse filtering with mathematical operations. From than the digital image quality degrade by blurring, noise, incorrect color balance and poor quality [4] which taken through image quality devices such as scanner, cameras and video recorder. To improve quality of digital images, various steps are required. The step involves i) Color correction to adjust the color of the image using color models or Color balancing method. K Murali Gopal and Ranjit Patnaik ii) Light illumination &Contrast enhancement to adjust the brightness. iii) Image smoothing by removing noise. iv) Image sharpening technique. Image enhancement is always a subjective evaluation means judgment is purely depend on viewer [5]. Image enhancement has many domains such as underwater vision [6], biomedical images [7], and outdoor vision [8]. II. BAD WEATHER The aim of getting knowledge about the weather is to design a weather free vision of surveillance and outdoor imaging [9, 10]. To design such a system is a challenging problem. The quality of an image or video in outdoor scene degraded due to the noise introduced by different environment efforts such as haze, fog, snow and rain [10]. The bad weather conditions can be of two types: 1. Steady or static Condition which introduce noise due to fog and haze in a regular pattern. 2. Dynamic condition which introduce noise due to rain and snow in an irregular pattern. Type and size of the particles and their concentration in space describe the noise in the image or video. [10] Asspecified in Table 1. Table 1 Weather condition & association type and size Condition Air Haze Fog Cloud Rain Particle Type Molecule Aerosol Water droplet Water droplet Water droplet Radius(μm) 10-4 10-2 – 1 1 - 10 1 - 10 102 - 104 A. HAZE Haze is traditionally an atmospheric phenomenon in which dust, smoke, and other dry particulates obscure the clarity of the sky. The term "haze", in meteorological literature, generally ijesird, Vol. IV, Issue VI, December 2017/212 International Journal of Engineering Science Invention Research & Development; Vol. IV, Issue VI, DECEMBER 2017 www.ijesird.com, E-ISSN: 2349-6185 is used to denote visibility-reducing aerosols of the wet type. Such aerosols commonly arise from complex chemical reactions that occur as sulfur dioxide gases emitted during combustion are converted into small droplets of sulphuric acid. The reactions are enhanced in the presence of sunlight, high relative humidity, and stagnant air flow. A small component of wet haze aerosols appear to be derived from compounds released by trees, such as terpenes. For all these reasons, wet haze tends to be primarily a warm-season phenomenon. Large areas of haze covering many thousands of kilometers may be produced under favorable conditions each summer. then becomes heavy enough to fall under gravity. Rain is a major component of the water cycle and is responsible for depositing most of the fresh water on the Earth. It provides suitable conditions for many types of ecosystems, as well as water for hydroelectric power plants and crop irrigation. Fig.3 Effect of Rain The overall picture of the bad weather image vision is classified as below: Weather image Fig.1 Effect of Haze B. FOG Fog consists of visible cloud water droplets or ice crystals suspended in the air at or near the Earth's surface. Fog can be considered a type of lowlying cloud and is heavily influenced by nearby bodies of water, topography, and wind conditions. In turn, fog has affected many human activities, such as shipping, travel, and warfare. Static Fog Dynamic Haze Rain Rain Streak Show Rain Drop ocs Unfocus Fig.4 Bad Weather classification Fig.2 Effect of Fog C. Rain Rain is liquid water in the form of droplets that have condensed from atmospheric water vapor and K Murali Gopal and Ranjit Patnaik III. LITERATURE SURVEY Single image restoration or enhancement of bad weather outdoor image mostly in rainy weather, it is an open research field with many problems. We focus in some of the related work on raindrop and ijesird, Vol. IV, Issue VI, December 2017/213 International Journal of Engineering Science Invention Research & Development; Vol. IV, Issue VI, DECEMBER 2017 www.ijesird.com, E-ISSN: 2349-6185 streaks remove from images in rainy environment as in Table 2 23 24 al Miyahara et. al Kurihata et. al Principle Component Analysis (eigendrops) Time-series 2008 2007 Table 2 Algorithms of detection and rain removal L. No. 1 Author Algorithm/ Technique Year Chen, DuanYuet al 2014 2 Pei, SooChang et al 3 Sun, ShaoHua et a Zhen g, Xianhui et al Framework guided image filter, low-frequency and a high frequency dictionary learning with sparse coding. Framework Merge Saturation and Visibility, High Pass Filter, Orientation Filter, Threshold. Incremental Dictionary learning-based method Using low frequency for single image and guided filter Neural network Adaptive nonlocal means filter. Guided image filter, then performing dictionary learning and sparse coding Guided filter Dictionary learning-based framework Framework based on morphological component analysis (MAC), bilateral filter, dictionary learning and sparse coding Refined guidance image Visual Salient Features Analyze an image into low frequency and high frequency via a bilateral filter and performs dictionary learning with sparse coding. Improved RIGSEC 2013 IV. RAINDROP CHARACTERISTICS Rain is a random shaped and sized water droplet traveling with high speed [11]. It is due to two reasons I. Initial differences in particle size II. Different rates of coalescences. The characteristics of raindrops are [12]  Edges that feature an outline of a raindrop  Blurry edges  Refraction of light  Consists of dark and bright region  Appears in circular form on windshield  Texture varies since the background varies  Causes blurring 2012 A. Bezier Curves Epipolar Geometry Intersection operation 2011 2011 2011 Intensity variation & contour verifications Support Vector Machine 2010 Improved RIGSEC 2009 RIGSEC 2009 Template matching 2008 4 5 6 7 8 9 10 Eigen , David et al Kim, JinHwan, et al Chen, DuanYu et al Xu, Jing et al Huan g, DeAn, et al LiWei Kang et.al 11 12 13 Jing Xu et. al Qi Wu et. al Fu, YuHsiang et. al 14 Sugimoto et. al Roser et. al Nomoto et. al Ching-Lin Yang Nashashibi et. al. Schwarxlmull er et. Al. Roser& Geiger Halimeh&Ros er [ Yamashita et. 15 16 17 18 19 20 21 22 K Murali Gopal and Ranjit Patnaik 2014 2014 2014 2013 2012 2012 2012 2012 2012 2012 2010 SIZE OF A RAINDROP The physical properties of rain have been extensively explored in atmospheric sciences and transportation. The size of a raindrop typically varies from 0.1 mm to 3.5 mm. B. SHAPE OF RAINDROP The shape of a drop can be expressed as a function of its size. Smaller raindrops are generally spherical in shape while larger drops resemble oblate spheroids. C. VELOCITY OF A RAINDROP ` During a normal rainfall, most of the drops are less than 1 mm in size. Hence, most raindrops are spherical. Therefore, this approximation in size is used to model the raindrops. As a drop falls through the atmosphere, it reaches a constant terminal velocity. V. CONCLUSION This article briefly describes preliminary study of the research fields in image processing and computer vision, more specific on image enhancement for weather degraded image. The main aim is to develop an algorithm to enhance image that can efficiently remove (raindrop) outdoor image within real-time processing. ijesird, Vol. IV, Issue VI, December 2017/214 International Journal of Engineering Science Invention Research & Development; Vol. IV, Issue VI, DECEMBER 2017 www.ijesird.com, E-ISSN: 2349-6185 6. 7. REFERENCES 8. 1. 2. 3. 4. 5. HUANG, T.S. K. AIZAWA. IMAGE PROCESSING: SOME CHALLENGING PROBLEMS. IN PROCEEDING OF NATIONAL ACADEMIC OF SCIENCES. 1993. WASHINGTON DC, USA. TWOGOODS, R.E., FUNDAMENTAL OF DIGITAL IMAGE PROCESSING, IN INTERNATIONAL SYMPOSIUM AND COURSE ON ELECTRONIC IMAGING IN MEDICINE1983: SAN ANTONIO, TEXAS. P. 1 - 19. HUANG, T.S., IMAGE ENHANCEMENT: A REVIEW. OPTICAL AND QUANTUM ELECTRONICS, 1969. 1(1): P. 49-59. ROSALINA, A.S., TAN, SAW KEOW, NURAINI, ABDUL RASHID, LIVE-CELL IMAGE ENHANCEMENT USING CENTRE WEIGHTED MEDIAN FILTER IN 11TH WSEAS INTERNATIONAL CONFERENCE ON COMPUTERS, A. NIKOLAOS, EDITOR 2007, WSEAS: CRETE ISLAND, GREECE. P. 382 - 385. WANG, D.C.C., A.H. VAGNUCCI, AND C.C. LI, DIGITAL IMAGE ENHANCEMENT: A SURVEY. COMPUTER VISION, GRAPHICS, AND IMAGE PROCESSING, 1983. 24(3): P. 363-381. AND K Murali Gopal and Ranjit Patnaik 9. 10. 11. 12. CELEBI, A.T. AND S. ERTURK, VISUAL ENHANCEMENT OF UNDERWATER IMAGES USING EMPIRICAL MODE DECOMPOSITION. EXPERT SYSTEMS WITH APPLICATIONS, 2011(0). ZIAEI, A., ET AL. A NOVEL APPROACH FOR CONTRAST ENHANCEMENT IN BIOMEDICAL IMAGES BASED ON HISTOGRAM EQUALIZATION. IN BIOMEDICAL ENGINEERING AND INFORMATICS, 2008. BMEI 2008. INTERNATIONAL CONFERENCE ON. 2008. NARASIMHAN, S.G. AND S.K. NAYAR, CONTRAST RESTORATION OF WEATHER DEGRADED IMAGES. IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE, 2003. 25(6): P. 713-724 COZMAN, F. AND E. KROTKOV. DEPTH FROM SCATTERING. IN COMPUTER VISION AND PATTERN RECOGNITION, 1997. PROCEEDINGS., 1997 IEEE COMPUTER SOCIETY CONFERENCE ON. 1997 NARASIMHAN, S.G. AND S.K. NAYAR, VISION AND THE ATMOSPHERE. INT. J. COMPUT. VISION, 2002. 48(3): P. 233- 254. GARG, K. AND S. NAYAR, VISION AND RAIN. INTERNATIONAL JOURNAL OF COMPUTER VISION, 2007. 75(1): P. 3-27 SCHWARZLMULLER, C., ET AL., A NOVEL SUPPORT VECTOR MACHINE CLASSIFICATION APPROACH INVOLVING CNN FOR RAINDROP DETECTION. ISAST TRANSACTIONS ON COMPUTERS AND INTELLIGENT VEHICLE SYSTEMS, 2010. 2(2): P. 52-65 ijesird, Vol. IV, Issue VI, December 2017/215