Prakhar Mohan

I was responsible for the developing and implementing computer vision/image processing algorithms for the company products and integrate them with the main product line and maintain the image processing module over time. My major contributions at GreyOrange were: 1) Empty Carriage Detection System (ECDS - V1.0): · Designed and Implemented a pattern matching algorithm to detect packages on the conveyor of a cross belt sortation system on an x86 based processor. · Acheived an accuracy of 98.3% with an average execution time of 7ms per image frame. · Published the results for the same at International Conference on Industrial Design Engineering (ICIDE), 2017. 2) Empty Carriage Detection System (ECDS - V2.0): · Implemented Yolo Object Detector to detect packages on the conveyor the a cross belt sortation system on NVIDIA Tegra TX2 - arm64 based mobile GPU. · Acheived an accuracy of 99.2% with an average execution time of 34ms per image frame. 3) Barcode Decoding and Pose Detection for an Autonomous Ground Vehicle: · Developed and implemented the algorithm to decode a barcode and estimate the pose of the AGV through it. · Reduced the error to less than 0.2 degree with a processing time of 9ms on x86 and ARM based processor.

I completed my Bachelor’s Thesis as a part of the internship program. Bachelor’s Thesis: Semi Automatic Segmentation of Objects in Images and Videos using Superpixels. Developed an algorithm to segment objects in images and videos using superpixels and temporal superpixel stitching