Majority of the high impedance faults (HIFs) occur through the ground and, due to the nonlinear variation in the fault signal, false detection rate is high. HIFs are accompanied with high frequency components and signal decomposition techniques are best suited to detect such faults. In an active distribution system, detection of HIF and its discrimination from capacitor switching, nonlinear switching and transformer energization is a challenging task. Also, the presence of noise in the signal complicates the detection task. A major gap in the detection process exists in the treatment of the negative impact of inverter based distributed generator. Taking the configurational and operational constraints into account, an Intrinsic Time Decomposition and Teager Kaiser Energy Operator (TKEO) based HIF detection approach is developed for distribution system. The first intrinsic mode function that contains the highest frequency component is extracted and applied to estimate the TKEO. The energy operator always surpasses the threshold range during HIF. The main advantages of this technique are that it has high immunity to noise, system unbalances and presence of harmonics and it does not rely on high current magnitude to detect the fault. Performance of the method is tested on several test cases.