In the glass fiber manufacturing process, the twisting process, as a crucial link between raw yarn production and weaving, has long been constrained by the challenge of controlling non-full-bore yarn, becoming the biggest obstacle to improving efficiency and reducing costs in the workshop. The residual yarn generated by the simultaneous operation of hundreds of spindles forces companies to adopt a passive model of "uniform reference, subsequent rework," with manual unwinding of each spindle resulting in cumbersome processes, low efficiency, high costs, uncontrollable quality risks, and extensive management. To address this industry pain point, a residual yarn detection system for twisting machines has emerged. Integrating grayscale recognition and distance detection technologies, it can monitor the amount of residual yarn in real time during twisting machine operation. Reaching a threshold triggers a single spindle shutdown, achieving in-process control of residual yarn and eliminating the risk of rework and residual yarn flowing downstream from the source, thus promoting the digital and intelligent upgrading of the glass fiber twisting process. The entire system consists of three core modules: the HD-CSJC residual yarn sensor, a signal converter, and a workshop server, offering stable networking and convenient deployment.