About distance measurement method for stray animal detection work

[SVinayak03]

Hello,
Really excited by looking your implementation, i need your help for my work. I working on stray animals detection and collision avoidance with yolov5.
I have obtained weights and trying for distance estimation. Can you please help me to know how your algorithm estimating distance with bounding box coordinates and standard scaling?
Please guide!

[HassanBinHaroon]

It's a pretty basic implementation of distance estimation (I have to mention it at the very start).

A simple neural network is being used for the distance estimation.

Kindly take a look at the concept of MONOCULAR DEPTH CUE.

The concept of MONOCULAR DEPTH CUE is the fundamental approach here.

The algorithm is the concatenation of two separate blocks. 1. Object Detection, 2. Distance Estimation.

The algorithm for object detection is a version of YOLO. While a dense network is for the distance estimation.

Kindly go through the README file thoroughly to reproduce the results and further improvements.

Moreover, I have to mention that there are other two repositories related to this work YOLOER_V1 and YOLOER_V2. You can look into all of those to get the insight.

Best,