We are doing a project in which we are detecting (using YOLOv4) runway debris using a fixed camera mounted at a pole on the side of the runway. We want to find out the position of the object with respect to the runway surface. Does anyone know about any algorithm or technique that will help us find out the position of the object with respect to runway dimensions?
Locate object in imageObject
Step 1: Read Images. Read the reference image containing the object of interest. Step 2: Detect Feature Points. Detect feature points in both images. ...
Step 3: Extract Feature Descriptors. ...
Step 4: Find Putative Point Matches. ...
Step 5: Locate the Object in the Scene Using Putative Matches. ...
Step 7: Detect Another Object. https://towardsdatascience.com/object-detection-with-10-lines-of-code-d6cb4d86f606
Related to your query, I suggest you follow https://leonardoaraujosantos.gitbook.io/artificial-inteligence/machine_learning/deep_learning/object_localization_and_detection
https://machinelearninhttps://
medium.com/@harman4422/object-localization-bd314d7e648f
gmastery.com/object-recognition-with-deep-learning/
Hassaan Zubair ,
Let us focus for a moment on the problem that you are trying to solve. The camera I am assuming operates under various conditions such as lighting variations(day/night), occlusions (shadows of nearby objects, and other weather conditions (fog, rain, etc.). From an image recognition problem it is not an easy problem to solve. Also, you are not constraining the debris, so it can be anything as of your initial post. You have also not mentioned: what is the runway for? Is it for cars, airplanes, motorcycles, etc. This variability has to be taken into account.
Now with the description of the problem ask yourself if a convolutional neural network is the best way to solve this problem.It is a supervised learning algorithm and this assumes that you have access to the right amount of data given the variability of your problem. I would also consider alternative algorithms besides a CNN for this problem and decompose the problem into various stages of cleanup before even doing debris recognition
Regards
Usually this is done using key point detection / localization algorithms like SIFT, SURF, KAZE, BRISK, HOG etc. all of which essentially look for change in image gradient magnitude and direction in local neighbourhoods.
- Badges
- Science topic
- Similar topics
- Biological Science
- Histology
- Fixatives