Improved Visual Hook Capturing and Tracking for Precision Hoisting of Tower Crane

Abstract EN

To maintain safe operation of the tower crane, it is important to monitor the activities of the hook system.

Visual monitoring and image recognition are the optimum methods for crane hook tracking and precision hoisting.

High real-time performance and low computation requirements are required for tower crane hook capturing and tracking system which is implemented on the embedded Advanced RISC Machines (ARM) processor or Microcontrol Unit (MCU).

Using the lift rope of a tower crane as the target object, a new high-performance hook tracking method suitble for ARM processor or MCU applications is presented.

The features of the lifting process are analyzed, and an improved progressive probabilistic Hough transform (IPPHT) algorithm is proposed which can reduce capturing time by up to 80%.

Combining color histogram with a binary search algorithm, an adaptive zooming method for precise hoisting is presented.

Using this method the optimum zoom scale can be achieved within a few iterations.