It is the process in converting radiated energy into temperature and advance radiometric cameras that filter puts out will allow you to gather the temperature data from every pixel within the scene and you can use this for post flight analysis using software like Flora tools for instance. This white image does not have this range on the right hand side that shows the temperature scale as shown in Fig. 3.
Fig. 3. Capabilities (Radiometry)
D. Covering Ground
Fig. 4. Capabilities Covering Ground
Each resolution and lens combination they put out a different field of view to capturing the symmetry. The basic concept here is that the wider field of view that you have the more you’ll see at a specific given time but you’ll also get less detail of the subject that you’re looking for. It was in the image . This can make locating a person quicker but also the possibility of missing them greater. The inverse is true with narrow field of view which gives you more specificity more detail but you can present a challenge in locating and staying locked on your target because it’s like looking through a soda straw. It’s important to consider basically for things the resolution of the camera the lens size and field of view the size of your target that you’re trying to identify and the distance between the camera and the target as shown in Fig. 4.
E. Ability to see in complete dark:
We’re particularly sensitive to infrared radiation. One of its most unique and useful characteristics that we will show simultaneous capture of visible imagery on the left part of the picture and thermal imagery on the right side of the picture as shown in Fig. 5.
Fig. 5. Clarity of view through thermal camera
So this is a night-time driving condition. There’s a reasonable amount of illumination. But if you watch the picture and pay attention you’ll notice a pedestrian starting to cross the road and through thermal video, we’re able to detect and classify the pedestrian crossing the road with the thermal camera significantly in advance of the visible camera as shown in Fig. 6 and Fig. 7.
Fig. 6. Clarity of view through thermal camera
Fig. 7. Pedestrian crossing the road
F. Not blinded by the sun
So we can see at night. We can see in complete darkness, one of the useful properties of thermal cameras. Another interesting and useful characteristic of thermal cameras is since we’re looking at heat and not light we’re not blinded when we’re driving into the sun as shown in Fig. 8.
Fig. 8. Thermal not blinded by the sun
Basically reduce the flare and glare that you see in the visible image here but fundamentally in a thermal image since we’re not sensitive to the light coming from the sun we don’t experience the same type of flare or glare that you see in visible camera.
G. Able to view in most type of fogs:
Another characteristic of long wave infrared thermal cameras is that we’re able to see exceptionally well in most types of fog as shown in Fig. 9 and Fig.10.
Fig. 9. Performs well in many types of fog (part1)
Fig. 10. Performs well in many types of fog (part2)
So we’re not bothered by the reflective light that’s impacting the performance of a physical camera. So you’ll be able to see in this scene if you pay attention to the thermal image and particularly along the sidewalk region here you’ll notice there are a number of pedestrians that you can easily see in the thermal image that you can’t see at all in the visible image. Here’s a pedestrian approaching the intersection that you’re able to detect very clearly thermally for example that you’re not able to see with the visible camera and the person who took this picture actually went bike riding along this path after the data collection and struggled to really see details and feel comfortable in that bike right afterwards. So thermal cameras do a great job in a lot of difficult environmental conditions like fog.
III. RESEARCH DESIGN AND METHODOLOGY
A. Data source credits : The data used for this study were collected by below sources:
1. [Las Vegas, United States based SPI Corp which gives thermal imaging frameworks to law authorization, border watch and country security]
2. [Based in Wilsonville, Oregon, United States, FLIR Systems Inc. FLIR is The Global Leader in the Design, Production, and Sales of Thermal Imaging Infrared Cameras]
3. [A multidisciplinary team at University of Klagenfurt and Lakeside Labs performs research on networked autonomous aerial systems.]
B. Technological design, workflow, tools , alogorithm ehhancements and methodology
TensorFlow can run on multiple CPU’s and GPU’s with optional Cuda for general purpose computing on graphics processing unit. And TensorFlow, its flexible architecture allows for easy deployment of computation across a variety of platforms which are the CPUs, GPU’s and TPU’s and from desktops to clusters and servers to mobile.
TensorFlow Context Diagram:
Now TensorFlow computations are expressed as state full data flow graphs. The name TensorFlow derives from the operations that such neural network performs on multi-dimensional data arrays which are known as the tensor. Now let’s see how object detection can be performed in tensor flow. So first of all we provide our input data which is a set of images and using tensor flow, we train our model. Now this model is strained using deep learning and the main object of this model is to extract the features as shown in Fig. 11.