A Study on the Development of Night Vision Thermal Camera Using Deep Learning

Technologies for pedestrian safety are increasingly emphasized by Automakers in advance of autonomous driving vehicles. A Night Vision System attached behind the front grille can reduce fatal accidents, especially during the nighttime, however, consumers may hesitate to adopt such systems on account of their high price. High-cost Germanium is used in commercial Night Vision System windows, and therefore replacing it with a cheaper infrared window material can lead to a more affordable system. To achieve this, Zinc Sulfide (ZnS), which has about 70% transmittance in the Long-Wavelength Infrared region of 8~12 μm, was selected for the window substrate material. In this study, we designed, fabricated and characterized a single layer cost-effective anti-reflection coating on a ZnS window substrate using Calcium Fluoride (CaF2). The CaF2 coating was fabricated by E-beam evaporation technique, with Quarter wavelength anti-reflection thickness (QAR). It was characterized by FT-IR, SEM and a thermal camera test module. We found that CaF2 both side coated the ZnS window and exhibited about 10~15% higher transmittance than the ZnS window substrate. In addition the CaF2 coating stably bonded to the ZnS substrate without any internal defects. A thermal camera based window test also showed better detection performance with the CaF2 Coating than a bare ZnS substrate window, which was calculated using the output voltage of the microbolometer thermal sensor.

Download Full-text

An Intelligent Microwave Oven with Thermal Imaging and Temperature Recommendation Using Deep Learning

Applied System Innovation ◽

10.3390/asi3010013 ◽

2020 ◽

Vol 3 (1) ◽

pp. 13 ◽

Cited By ~ 2

Author(s):

Tareq Khan

Keyword(s):

Deep Learning ◽

Real Time ◽

Thermal Imaging ◽

Closed Loop ◽

Thermal Image ◽

Microwave Oven ◽

Closed Loop Control ◽

Target Temperature ◽

Thermal Camera ◽

Loop Control

Whenever food in a microwave oven is heated, the user estimates the time to heat. This estimation can be incorrect, leading the food to be too hot or still cold. In this research, an intelligent microwave oven is designed. After the food is put into the microwave oven and the door is closed, it captures the image of the food, classifies the image and then suggests the food’s target temperature by learning from previous experiences, so the user does not have to recall the target food temperature each time the same food is warmed. The temperature of the food is measured using a thermal camera. The proposed microwave incorporates a display to show a real-time colored thermal image of the food. The microwave automatically stops the heating when the temperature of the food hits the target temperature using closed-loop control. The deep learning-based image classifier gradually learns the type of foods that are consumed in that household and becomes smarter in temperature recommendation. The system can classify and recommend target temperature with 93% accuracy. A prototype is developed using a microcontroller-based system and successfully tested.

Download Full-text

Infrared visible color night vision image fusion based on deep learning

AI and Optical Data Sciences II ◽

10.1117/12.2592791 ◽

2021 ◽

Author(s):

Yan Zou ◽

Linfei Zhang ◽

Chengqian Liu ◽

Bowen Wang ◽

Yan Hu ◽

...

Keyword(s):

Deep Learning ◽

Image Fusion ◽

Night Vision ◽

Color Night Vision

Download Full-text

Deep Learning Thermal Image Translation for Night Vision Perception

ACM Transactions on Intelligent Systems and Technology ◽

10.1145/3426239 ◽

2020 ◽

Vol 12 (1) ◽

pp. 1-18

Author(s):

Shuo Liu ◽

Mingliang Gao ◽

Vijay John ◽

Zheng Liu ◽

Erik Blasch

Keyword(s):

Deep Learning ◽

Thermal Image ◽

Night Vision ◽

Image Translation

Download Full-text

Deep Learning Based Super Resolution and Classification Applications for Neonatal Thermal Images

Traitement du signal ◽

10.18280/ts.380511 ◽

2021 ◽

Vol 38 (5) ◽

pp. 1361-1368

Author(s):

Fatih M. Senalp ◽

Murat Ceylan

Keyword(s):

Deep Learning ◽

High Resolution ◽

Thermal Imaging ◽

Similarity Index ◽

Super Resolution ◽

Ground Truth ◽

Structural Similarity ◽

Generative Adversarial Networks ◽

Thermal Camera ◽

Thermal Images

The thermal camera systems can be used in all kinds of applications that require the detection of heat change, but thermal imaging systems are highly costly systems. In recent years, developments in the field of deep learning have increased the success by obtaining quality results compared to traditional methods. In this paper, thermal images of neonates (healthy - unhealthy) obtained from a high-resolution thermal camera were used and these images were evaluated as high resolution (ground truth) images. Later, these thermal images were downscaled at 1/2, 1/4, 1/8 ratios, and three different datasets consisting of low-resolution images in different sizes were obtained. In this way, super-resolution applications have been carried out on the deep network model developed based on generative adversarial networks (GAN) by using three different datasets. The successful performance of the results was evaluated with PSNR (peak signal to noise ratio) and SSIM (structural similarity index measure). In addition, healthy - unhealthy classification application was carried out by means of a classifier network developed based on convolutional neural networks (CNN) to evaluate the super-resolution images obtained using different datasets. The obtained results show the importance of combining medical thermal imaging with super-resolution methods.

Download Full-text

Visible and Thermal Camera-Based Jaywalking Estimation Using a Hierarchical Deep Learning Framework

Computer Vision – ACCV 2020 Workshops - Lecture Notes in Computer Science ◽

10.1007/978-3-030-69756-3_9 ◽

2021 ◽

pp. 123-135

Author(s):

Vijay John ◽

Ali Boyali ◽

Simon Thompson ◽

Annamalai Lakshmanan ◽

Seiichi Mita

Keyword(s):

Deep Learning ◽

Thermal Camera ◽

Learning Framework

Download Full-text

Human action recognition for night vision using temporal templates with infrared thermal camera

2013 10th International Conference on Ubiquitous Robots and Ambient Intelligence (URAI) ◽

10.1109/urai.2013.6677407 ◽

2013 ◽

Cited By ~ 4

Author(s):

Hyukmin Eum ◽

Jeisung Lee ◽

Changyong Yoon ◽

Mignon Park

Keyword(s):

Action Recognition ◽

Human Action Recognition ◽

Human Action ◽

Night Vision ◽

Thermal Camera

Download Full-text

Multiple Linear Regression and Deep Learning in Body Temperature Detection and Mask Detection

IT JOURNAL RESEARCH AND DEVELOPMENT ◽

10.25299/itjrd.2022.7424 ◽

2021 ◽

pp. 109-121

Author(s):

Faisal Najib Abdullah ◽

Mohamad Nurkamal Fauzan ◽

Noviana Riza

Keyword(s):

Deep Learning ◽

Body Temperature ◽

Linear Regression ◽

Multiple Linear Regression ◽

Raspberry Pi ◽

Thermal Camera ◽

The Public ◽

Temperature Detection ◽

Distance Data ◽

The Government

In this new normal era, many activities began to operate again, such as offices, malls, etc. This creates a potential mass crowd. The public must follow health protocols as recommended by the government, including wearing masks and checking the temperature to anticipate the spread of the coronavirus. This study tested a tool that included image processing and artificial intelligence to help implement health protocols as recommended by the government. This tool connects Raspberry PI, Thermal Camera (amg8833), Pi Camera, an ultrasonic sensor with Multiple Linear Regression and Deep Learning algorithms. The purpose of this tool is to detect body temperature and detect the use of masks. The system will check on the pi camera frame whether the person is wearing a mask or not. The system is trained using the Deep Learning method to detect the use of masks. The system will check the temperature of the human body and the distance between humans and the tool. Temperature and distance data are entered in multiple linear regression formulas to get more accurate results. The processed results of the system will be displayed on the monitor screen if detected using a mask and the normal temperature will be green and if it is not detected it will be red and give a warning sound. The data is sent to the server and displayed via the web. We found that this tool succeeded in detecting body temperature within a distance of 1 to 3 meters with an accuracy of 99.49%, detecting people using masks with an accuracy of 94.71%, and detecting people not wearing masks with an accuracy of 97.7%.

Download Full-text