Weekly Progress Report #1

Weekly Progress Report #1
Adnan Khan and Noah Borel
10/25/15


Progress:
1. Overall System Diagram and Plan Created:
 -   Both the Vision System team and the Interception System team have created a system diagram of the major aspects of the entire drone defense system. An order of events has been made according to the system diagram and tasks have been assigned. 
 -   We will be using a normal webcam for the beginning to just get familiar with how the drone detection will work. Then, after we can track with a webcam, we will look into the Amcrest cameras.

2. OpenCV:
 -   Research has been done into downloading OpenCV and connecting it with a development software.
Commonly used development software: Microsoft Visual Studio, Python
Links: 1. https://www.visualstudio.com/en-us/visual-studio-homepage-vs.aspx
           2. https://www.python.org

3. Drone Detection:
 -   Research has been made into what will be done once the development software is up and running and how the drone will be tracked.
 -   RGB (red, green, blue) colored image that is taken is converted into HSV (hue, saturation, value) to get more distinction in the resulting image.
 -   HSV image is altered with thresholds to isolate the drone. If drone is a certain distinct HSV color, then thresholds can help to isolate it so that the range of HSV of the drone is only only seen. Lastly, this image is converted to a binary image of white and black where the thresholds help to only reveal the drone.
Links: 1.  https://www.youtube.com/watch?v=bSeFrPrqZ2A
           2.  http://computer-vision-tutorials.blogspot.com/2014/03/opencv-tutorials-color-detection-object.html

















4. Compound Eye:
 -   Found book on Compound Eyes: "The Physiology of the Compound Eyes of Insects and Crustaceans" By Sigmund Exner. Even though the original book was written in 1891, it includes a deep study of the compound eye. The text has much more explanation than in a dissertation from a college student.

Problems:
 -   OpenCV: There are many options/paths that could be taken with which development software could be used. Some are able to use Mac OSX while others cannot. The options must be carefully looked through.

 -   Compound Eye: There is a lot of complicated math involved in the analysis of the compound eye. There are many variables and formulas that relate to the optics of the eye.

Plan:
 -   OpenCV: Do a study of the various development softwares and which one would be best considering which computers we will be using. Learn the process that must go into each software to link OpenCV.

 -   Drone Detection: Once the OpenCV step is complete, the coding for the drone detection can be implemented.

 -   Compound Eye: With any math or diagrams that seem to complicated, I will consult with Mr. Lin to better understand them.

Camera Overviews

THETA 

Products: Ricoh Theta, Ricoh Theta m15, Ricoh Theta S

For all downloadable software is needed to view the videos or images recorded on a Mac or PC.

Ricoh Theta:
- Takes only still images.
- Can take 360° images with one shot.
- Can be controlled wirelessly with a smartphone.
- Images can't be viewed live, need to be separately uploaded onto a PC, Mac or smartphone.
- Can't be used when charging.

Ricoh Theta m15:
- Can take videos as well as image.
- Videos and images needed to be transferred to PC, Mac or a smartphone to be viewed.
- Can take 360° images and videos.
- Can't be used when charging.
- Can be controlled by a smartphone wirelessly.
- Can only shoot video for 5 minutes

Ricoh Theta S:
- Higher quality video and images (1080p videos at 30fps).
- Can take 360° images and videos.
- Can take live video at 10fps and broadcast it to smartphone.
- Can shoot videos for up to 25 minutes.
- Can't shoot videos while charging.
- Can be controlled by a smartphone.

Ideas for Eye Structure

Ideas for Eye Structure
10/18/15
Adnan Khan


1. Single aperture camera mounted facing upward. Custom made glass attachments made to extend field of view and bend incoming light to the single lens. Each glass piece is angled precisely to have the image from different angles be mirrored into the lens. The received images would be in quadrants or in divided sections with each section found in a different square in the array.
2. Mounted omnidirectional camera. The camera revolves around and captures images of the full semi-sphere field of view piece by piece. The drone could have a unique color that would be detected. The camera would be in a certain order of image capturing, but when the unique color is detected the camera will stop.
3. A single video-recording camera. This camera could be mounted on a device that could move it in a path to have it record all in the semi-sphere field of view. The camera movement will not stop when the drone is detected, but the point on the camera's path where the drone is detected will be recorded.
4. Stationary Ricoh style camera with two opposite facing cameras. Camera could be rotated so that the wide field of view cameras are able to see where there would be blind spots.
5. Three angled single aperture cameras that face upward and outward. The three single aperture cameras act as three ommatidia with image capturing of only that one field of view. The camera would be stationary.

Object Detection Research - Noah Borel

1. A Survey on Approaches of Object Detection (2013),  S. Shantaiya, K. Verma, and K. Mehta.

Overview:

- Object detection using video has progressed rapidly in recent years.

- In most cases the focus has been on human motion and behavior.

Introduction:

- Video detection is starting to be used in detection of pedestrians, monitoring car traffic and identifying strange behavior near ATMs.

- Video detection has a huge interest because of it's potential uses in security.

Object Detection Approaches:

- High quality but inexpensive video cameras as well as high-powered computers are used in object detection.

- The detection of moving objects in video is a key step in overall object detection.

Feature Based Object Detection:

- In this type of object detection the object is usually detected using shape, size or color.

Shape:

- Shape based object detection is very complex due to having to segment the object from the rest of the picture.

- It becomes increasingly harder when there are multiple objects in the picture with different shadings and light levels.

- A method for detection of athletes in game is the PCA-HOG. This involves transforming the athletes into Histograms of Oriented Gradient and then applying Principal Component Analysis to them.

- A limitation of this particular method is that it is difficult to detect multiple objects. 

- Another method is by using variable resolution double-level contours. This involves using a low resolution image to detect the object's edges and get the outline of the object. Then the high resolution image is used to in a process where the outline reduces until the actual object outline is found.

- One other method for shape based detection is to automatically differentiate between the foreground and background of the image and to detect the object based on the assumption that it has different geometrical features than the background.

Color:

- Color detection requires low computational power relative to other methods.

- A method for object detection using color involves the analyzing of RGB color images. A downside to this is that detection fails if the object is too small.

- Another color based detection involves using color histograms. This follows the idea of extracting HOG features with pixel colors inside being analyzed. A drawback of this method is that it can't work if the background is a similar color to the object.

Template Based Detection

- This requires a template of the object being detected to be available.

- Detection is done by matching the features of the template to the image being observed.

- The two types of template detection are fixed and deformable template matching.

- Fixed template matching is very useful if the object's shape does not vary with the change in viewing angles.

- Deformable template matching is more suitable in most cases of detection because objects tend to not have the same shape from different angles of view.

Motion Based Detection

- This method relies on changes over pixel or block levels. 

- The first stationary frame is used as reference background frame. The preceding frames with the object in them are then compared to the reference frame and detect the object. This method requires the object to be moving continuously and not be in the image before being detected.